Cisco Desktop Virtualization Solution for EMC VSPEX with VMware Horizon View 5.3 for 2000 Desktops
Available Languages
Table of Contents
About Cisco Validated Design (CVD) Program Cisco Solutions for EMC VSPEX End User Computing: VMware Horizon View 5.3
Benefits of Cisco Unified Computing System
Benefits of EMC VNX Family of Storage Controllers
Benefits of VMware vSphere ESXi 5.5
Benefits of VMware Horizon View 5.3
Easily Control and Manage Desktops
Deliver a Better Desktop Experience
Automate Desktop Operations Management
Cisco Unified Computing System (UCS)
Cisco Unified Computing System Components
Cisco UCS 2200 Series IO Module
Cisco UCS B200 M3 Blade Server
What\qs New in Horizon View 5.3
VMware Horizon View 5.3 Hosted Virtual Desktop (HVD) Overview
Modular Virtual Desktop Infrastructure Technical Overview
Understanding Desktop User Groups
Understanding Applications and Data
Project Planning and Solution Sizing Sample Questions
The Solution: A Unified, Pre-Tested and Validated Infrastructure
Cisco Networking Infrastructure
Cisco Nexus 5500 Series Feature Highlights
Cisco Nexus 1000V Feature Highlights
Architecture and Design of Horizon View 5.3 on Cisco Unified Computing System and EMC VNX Storage
Hosted Virtual Desktop (HVD) Design Fundamentals
Choose a User Profile Management System
Accessing USB Devices Connected to the End Point
Designing a VMware Horizon View 5.3 Deployment
Determine Desktop Pools Required
Hosted Virtual Desktop Infrastructure Design
Desktop Delivery Base Image Creation
Microsoft Windows 7 Golden Image Creation
Create base Windows SP1 Virtual Machine
Optimization of base windows 7 SP1 virtual machine
Install View 5.3 Virtual Desktop Agent software
Perform Additional View 5.3 Configuration
Cisco Unified Computing System Configuration
QoS and CoS in Cisco Unified Computing System
Cisco UCS System Class Configuration
Enable QOS on the Cisco Unified Computing System
Nexus 1000V Configuration in L3 mode.
Configuring Boot from SAN Overview
SAN Configuration on Nexus 5548UP
Configuring Boot from SAN on EMC VNX
SAN Configuration on Cisco UCS Manager
EMC VNX5600 Storage Configuration
EMC Storage Configuration for VMware View
Example EMC Boot LUN Configuration
Cisco UCS Manager Configuration for VMware ESXi 5.5
Installing and Configuring ESXi 5.5
ESXi 5.5 Cluster Configuration
Infra-Cluster Infrastructure Cluster
Installing and configuring VMware View 5.3
Create SQL Databases for View 5.3
Install View Manager and components
View Administrator Configuration
Install SSL Certificate for View Connection and Replica Server
Install View Client on End Points
Configure the View 5.3 Hosts and Storage
Configure the View Desktop Pools and Options.
Cisco UCS Test Configuration for Single Blade Scalability
Cisco UCS Configuration for Two Chassis – Fourteen Blade Test
Testing Methodology and Success Criteria
User Workload Simulation LoginVSI from Login Consultants
Cisco UCS Test Configuration for Single-Server Scalability Test Results
Cisco UCS Test Configuration for 2000 Desktop Two-Cluster Scalability Test Results
Scalability Considerations and Guidelines
Cisco UCS System Configuration Considerations
VMware View 5.3 Considerations
EMC VNX Storage Guidelines for Horizon View 5.3 Provisioned Virtual Machines
VMware ESXi 5.5 Guidelines for Virtual Desktop Infrastructure
VMware View Reference Documents
About Cisco Validated Design (CVD) Program
- The CVD program consists of systems and solutions designed, tested, and documented to facilitate faster, more reliable, and more predictable customer deployments. For more information visit:
- http://www.cisco.com/go/designzone
- ALL DESIGNS, SPECIFICATIONS, STATEMENTS, INFORMATION, AND RECOMMENDATIONS (COLLECTIVELY, "DESIGNS") IN THIS MANUAL ARE PRESENTED "AS IS," WITH ALL FAULTS. CISCO AND ITS SUPPLIERS DISCLAIM ALL WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE WARRANTY OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT OR ARISING FROM A COURSE OF DEALING, USAGE, OR TRADE PRACTICE. IN NO EVENT SHALL CISCO OR ITS SUPPLIERS BE LIABLE FOR ANY INDIRECT, SPECIAL, CONSEQUENTIAL, OR INCIDENTAL DAMAGES, INCLUDING, WITHOUT LIMITATION, LOST PROFITS OR LOSS OR DAMAGE TO DATA ARISING OUT OF THE USE OR INABILITY TO USE THE DESIGNS, EVEN IF CISCO OR ITS SUPPLIERS HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.
- THE DESIGNS ARE SUBJECT TO CHANGE WITHOUT NOTICE. USERS ARE SOLELY RESPONSIBLE FOR THEIR APPLICATION OF THE DESIGNS. THE DESIGNS DO NOT CONSTITUTE THE TECHNICAL OR OTHER PROFESSIONAL ADVICE OF CISCO, ITS SUPPLIERS OR PARTNERS. USERS SHOULD CONSULT THEIR OWN TECHNICAL ADVISORS BEFORE IMPLEMENTING THE DESIGNS. RESULTS MAY VARY DEPENDING ON FACTORS NOT TESTED BY CISCO.
- The Cisco implementation of TCP header compression is an adaptation of a program developed by the University of California, Berkeley (UCB) as part of UCB’s public domain version of the UNIX operating system. All rights reserved. Copyright © 1981, Regents of the University of California.
- Cisco and the Cisco Logo are trademarks of Cisco Systems, Inc. and/or its affiliates in the U.S. and other countries. A listing of Cisco's trademarks can be found at http://www.cisco.com/go/trademarks . Third party trademarks mentioned are the property of their respective owners. The use of the word partner does not imply a partnership relationship between Cisco and any other company. (1005R)
- Any Internet Protocol (IP) addresses and phone numbers used in this document are not intended to be actual addresses and phone numbers. Any examples, command display output, network topology diagrams, and other figures included in the document are shown for illustrative purposes only. Any use of actual IP addresses or phone numbers in illustrative content is unintentional and coincidental.
- © 2014 Cisco Systems, Inc. All rights reserved.
Overview
Industry trends indicate a vast data center transformation toward shared infrastructures. Enterprise customers are moving away from silos of information and toward shared infrastructures, to virtualized environments, and eventually to the cloud to increase agility and reduce costs.
VSPEX is a flexible architecture solution that allows IT to quickly and consistently deploy a virtualized infrastructure to consolidate servers and applications that help reduce capital and operational expenses.
EMC VSPEX End User Computing solutions for VMware Horizon View 5.2 allows a more strategic approach. Technology partners working with EMC deliver VSPEX Proven Infrastructure consisting of best-in-class technologies built as easy to manage complete virtualization solutions. With the leading virtualization platform (VMware vSphere) together with VMware EUC software, Cisco compute and networking, and EMC storage and backup technologies it’s possible simplify IT, centrally manage desktops, applications, and end user data.
This Cisco Solution for EMC VSPEX End User Computing reports the results of a study evaluating the scalability of VMware Horizon View 5.3 environment on Cisco UCS B-Series B200-M3 Blade servers running on VMware ESXi 5.5 hypervisor software connected to an EMC VNX 5600 Storage Array. We utilize second and third generation Unified Computing System hardware and software. We provide best practice recommendations and sizing guidelines for large scale customer deployments of VMware Horizon View 5.3 on the Cisco Unified Computing System.
Solution Component Benefits
Each of the components of the overall solution materially contributes to the value of functional design contained in this document. Reduces Risk, Interoperability concerns, sizing issues, performance concerns and enables you to realize a return on your investment sooner.
Benefits of Cisco Unified Computing System
Cisco Unified Computing System™ is the first converged data center platform that combines industry-standard, x86-architecture servers with networking and storage access into a single converged system. The system is entirely programmable using unified, model-based management to simplify and speed deployment of enterprise-class applications and services running in bare-metal, virtualized, and cloud computing environments.
Benefits of the Unified Computing System include:
- Third generation B-Series blade servers for infrastructure and virtual workload hosting
- Third generation C-Series rack-mount servers for infrastructure and virtual workload Hosting
- 6200 Series second generation fabric interconnects provide unified blade, network and storage connectivity
- 5108 Blade Chassis provide the perfect environment for multi-server type, multi-purpose workloads in a single containment
- Converged, simplified architecture drives increased IT productivity
- Cisco UCS Manager results in flexible, agile, high performance, self-integrating information technology infrastructure with faster ROI
- Fabric Extender technology reduces the number of system components to purchase, configure and maintain
- Standards-based, high bandwidth, low latency virtualization-aware unified fabric delivers high density, excellent virtual desktop user-experience
Benefits of Nexus 5548UP
The Cisco Nexus 5548UP Switch delivers innovative architectural flexibility, infrastructure simplicity, and business agility, with support for networking standards. For traditional, virtualized, unified, and high-performance computing (HPC) environments, it offers a long list of IT and business advantages, including:
- Unified ports that support traditional Ethernet, Fibre Channel (FC), and Fibre Channel over Ethernet (FCoE)
- Synchronizes system clocks with accuracy of less than one microsecond, based on IEEE 1588
- Offers converged Fabric extensibility, based on emerging standard IEEE 802.1BR, with Fabric Extender (FEX) Technology portfolio, including:
- Nexus 1000V Virtual Distributed Switch
- Cisco Nexus 2000 FEX
- Adapter FEX
- VM-FEX
- Common high-density, high-performance, data-center-class, fixed-form-factor platform
- Consolidates LAN and storage
- Supports any transport over an Ethernet-based fabric, including Layer 2 and Layer 3 traffic
- Supports storage traffic, including iSCSI, NAS, FC, RoE, and IBoE
- Reduces management points with FEX Technology
Benefits of EMC VNX Family of Storage Controllers
The EMC VNX Family delivers industry leading innovation and enterprise capabilities for file, block, and object storage in a scalable, easy-to-use solution. This next-generation storage platform combines powerful and flexible hardware with advanced efficiency, management, and protection software to meet the demanding needs of today’s enterprises.
All of this is available in a choice of systems ranging from affordable entry-level solutions to high performance, petabyte-capacity configurations servicing the most demanding application requirements. The VNX family includes the VNXe Series, purpose-built for the IT generalist in smaller environments, and the VNX Series, designed to meet the high-performance, high scalability, requirements of mid size and large enterprises.
VNX Series - Simple, Efficient, Powerful
A robust platform for consolidation of legacy block storage, file-servers, and direct-attached application storage, the VNX series enables organizations to dynamically grow, share, and cost-effectively manage multi-protocol file systems and multi-protocol block storage access. The VNX Operating environment enables Microsoft Windows and Linux/UNIX clients to share files in multi-protocol (NFS and CIFS) environments. At the same time it supports iSCSI, Fiber Channel, and FCoE access for high bandwidth and latency-sensitive block applications. The combination of EMC Atmos Virtual Edition software and VNX storage supports object-based storage and enables customers to manage web applications from EMC Unisphere. The VNX series next generation storage platform is powered by Intel quad-core Xeon 5600 series with a 6 –GB/s SAS drive back-end and delivers demonstrable performance improvements over the previous generation mid-tier storage:
Benefits of VMware vSphere ESXi 5.5
As virtualization is now a critical component to an overall IT strategy, it is important to choose the right vendor. VMware is the leading business virtualization infrastructure provider, offering the most trusted and reliable platform for building private clouds and federating to public clouds.
Find out how only VMware delivers on the core requirements for a business virtualization infrastructure solution:
- Is built on a robust, reliable foundation
- Delivers a complete virtualization platform from desktop through the data center out to the public cloud Provides the most comprehensive virtualization and cloud management
- Integrates with your overall IT infrastructure
- Is proven over with 350,000 customers
- VMware delivers while providing Low total-cost-of-ownership (TCO)
See what is new in vSphere 5.5 by visiting the following URL:
http://www.vmware.com/files/pdf/vsphere/VMware-vSphere-Platform-Whats-New.pdf .
The Release Notes are available here: https://www.vmware.com/support/vsphere5/doc/vsphere-esx-vcenter-server-55-release-notes.html
Benefits of VMware Horizon View 5.3
Deliver rich, personalized virtual desktops as a managed service from a virtualization platform built to deliver the entire desktop, including the operating system, applications and data. With Horizon View, desktop administrators virtualize the operating system, applications, and user data and deliver modern desktops to end-users. Get centralized automated management of these components for increased control and cost savings. Improve business agility while providing a flexible high performance desktop experience with VMware Horizon View, desktop administrators virtualize the operating system, applications, and user data and deliver modern desktops to end-users, across a variety of network conditions.
Deliver Business Agility
Bring the agility and availability of cloud computing to the desktop and applications with Horizon View. Built on VMware vSphere, Horizon View delivers desktops from a single integrated platform as part of your cloud services.
Dynamically allocate resources to enable highly responsive environment to end users. Scale up and down desktop services on demand to quickly meet changing business needs and pro actively protect against planned and unplanned downtime. Run your desktops as business critical services for your workforce.
Easily Control and Manage Desktops
Increase control of desktops, applications and data by delivering and managing them as centralized services.
A single, powerful administrative console provides oversight of desktop services while enabling IT to simply execute previously cumbersome tasks like provisioning, updates and patches. Easily apply policies, quickly enable and disable users all from a centralized console for optimal business response. Free up time from maintenance for technology innovation.
Deliver a Better Desktop Experience
Unlike traditional PCs, View desktops are not tied to the physical computer. Instead, they reside in your cloud and end-users can access their View desktop when needed.
Horizon View with PCoIP delivers the richest, most flexible and adaptive experience for end-users around the world in a variety of network conditions. Business happens everywhere, online or offline, desktops or mobile devices, LAN or WAN, Horizon View delivers maximum workplace productivity.
Automate Desktop Operations Management
VMware vCenter Operations Manager for View allows administrators to gain insight into desktop and infrastructure performance, quickly pinpoint and troubleshoot issues. Administrators can optimize resource utilization, and pro actively manage the desktop environment through the management dashboards. vCenter Operations Manager for View is an optional add-on for Horizon View customers. You can also leverage PCoIP Extension Services to collect Horizon View statistics into your existing WMI tool.
Built-in Security
Maintain control over data and intellectual property by keeping it secure in the data center. Encrypted protocol traffic provides secure end-users access virtual desktops inside or outside of the corporate network. Integration with vShield Endpoint enables offloaded and centralized anti-virus and anti-malware (AV) solutions. This integration helps to eliminate agent sprawl and AV storm issues while minimizing the risk of malware infection and simplifying AV administration. VMware View also supports integration with RSA SecureID for 2-factor authentication requirements.
Audience
This document describes the architecture and deployment procedures of an infrastructure comprised of Cisco, EMC, and VMware hypervisor and desktop virtualization products. The intended audience of this document includes, but is not limited to, sales engineers, field consultants, professional services, IT managers, partner engineering, and customers who want to deploy the solution described in this document.
Summary of Main Findings
The combination of technologies from Cisco Systems Inc, VMware and EMC produced a highly efficient, robust and scalable Desktop Virtualization (DV) infrastructure for a hosted virtual desktop deployment. Key components of the solution included:
- The combined power of the Unified Computing System, Nexus switching and EMC storage hardware with VMware ESXi 5.5, and VMware Horizon View 5.3 software produces a high density per blade and per chassis Virtual Desktop delivery system.
- B200 M3 half-width blade with dual 10-core processors and 384GB of memory running at 1333 MHz supports 30% more virtual desktop workloads than the previously studied full width blade using a new medium workload with flash. In addition, density achieved with Horizon View 5.3 is equivalent to a prior study on the same platform with View 5.1 Update 2.
- The study design based on two Unified Computing System chassis, each with seven B200-M3 blades, each with dual 10-core processors and 384GB of memory running at 1333 MHz and a Cisco VIC 1240 converged network adapter supports 2000 virtual desktop workloads running the new medium workload with flash providing outstanding End User Experience with average response times under 1.75 seconds at full load.
- We were able to boot the full complement of 2000 virtual desktops (ready to login) in under 16 minutes.
- We were able to ramp up (log in and start workloads) to steady state with all 2000 users running a knowledge worker workload with flash in 30 minutes without pegging the processor, exhausting memory or storage subsystems.
- Our design provides N+1 server fault tolerance for the 2000 virtual desktop system, making the design fully fault tolerant from end to end.
- Compared to previous studies with full width blades, the rack space required to support 2000 users was reduced from 30 Rack Units to 12 Rack Units.
- Pure Virtualization: We continue to present a validated design that is 100% virtualized on ESXi 5.5. All of the Windows 7 SP1 virtual desktops and supporting infrastructure components, including vCenter, Active Directory, Profile Servers, SQL Servers, and Horizon View 5.3 components were hosted as virtual servers.
- We maintain our industry leadership with our new Cisco UCS Manager 2.2(1b) software that makes scaling simple, consistency guaranteed and maintenance simple.
- Our 10G unified fabric story gets additional validation on second generation 6200 Series Fabric Interconnects and second generation Nexus 5500 Series access switches as we run more challenging workload testing, maintaining unsurpassed user response times.
- EMC’s VNX5600 system provides storage consolidation and outstanding efficiency. Both block and file based storage resources are available on a single system, utilizing EMC Fast Cache technology.
- VMware Horizon View 5.3 with the Sparse Virtual Disk feature used for floating assignment linked clones provided better disk performance and space efficiency.
Architecture
Hardware Deployed
The architecture deployed is highly modular. While each customer’s environment might vary in its exact configuration, once the reference architecture contained in this document is built, it can easily be scaled as requirements and demands change. This includes scaling both up (adding additional resources within a UCS Domain) and out (adding additional UCS Domains and VNX Storage arrays).
The 2000 User Horizon View 5.3 solution includes Cisco networking, Cisco UCS and EMC storage, all of which fits in two data center racks (one for the EMC VNX and one for the Cisco networking and Cisco UCS gear.) In fact, there is adequate rack space in the Cisco rack to add blades and chassis to support an additional 4000 users
This document details the deployment of VMware Horizon View 5.3 floating assignment linked clones on VMware ESXi 5.5. Cisco Nexus 1000V distributed switch manages the two VMware Clusters hosting the virtual desktops, insuring end to end Quality of Service and ease of management by the network team.
Figure 1 VMware Horizon View 5.3 2000 User Hardware Components
The reference configuration includes:
- Two Cisco Nexus 5548UP switches with 16-universal port Expansion Modules (Optional)
- Two Cisco UCS 6248 Series Fabric Interconnects with UCS 6200 16-universal port Expansion Modules (Optional)
- Two Cisco UCS 5108 Blade Server Chassis with two 2204XP IO Modules per chassis
- Fourteen Cisco UCS B200 M3 Blade Servers with Intel E5-2680v2 processors, 384GB RAM, and VIC 1240 mezzanine cards for Horizon View 5.3 virtual desktops (providing N+1 Server fault tolerance for the system)
- One EMC VNX5600 dual controller storage system for HA
- Two Cisco UCS B200 M3 Blade servers with Intel E5-2650v2 processors, 128 GB RAM, and VIC 1240 mezzanine card for infrastructure (not shown in the drawing above)
The EMC VNX5600 disk shelf, disk and Fast Cache configurations are detailed in Section 5.4 Storage Architecture Design later in this document.
Configuration Guidelines
The 2000 User Horizon View 5.3 solution described in this document provides details for configuring a fully redundant, highly-available configuration. Configuration guidelines are provided that refer to which redundant component is being configured with each step, whether that be A or B. For example, SP A and SP B are used to identify the two EMC VNX storage controllers that are provisioned with this document while Nexus A and Nexus B identify the pair of Cisco Nexus switches that are configured. The Cisco UCS Fabric Interconnects are configured similarly.
This document is intended to allow the reader to configure the VMware Horizon View 5.3 customer environment as stand-alone solution.
VLANs
For the 2000 User Horizon View 5.3 solution, we utilized VLANs to isolate and apply access strategies to various types of network traffic. Table 2 details the VLANs used in this study
VMware Clusters
We utilized four VMware Clusters to support the solution and testing environment:
- Infrastructure Cluster (vCenter, Active Directory, DNS, DHCP, SQL Clusters, VMware View Connection Servers, View Composer, and Nexus 1000V Virtual Switch Manager appliances, etc.)
- VDI Cluster (Windows 7 SP1 32-bit pooled virtual desktops; 1000 virtual machines per replica as per VMware best practices recommended Horizon View 5.3 desktop cluster density.)
- Launcher Cluster (The Login Consultants Login VSI launcher infrastructure was hosted on the same UCS Domain sharing switching, but running on local storage.)
Infrastructure Components
This section describes the entire infrastructure components used in the VSPEX solution outlined in this study.
Cisco Unified Computing System (UCS)
Cisco UCS is a set of pre-integrated data center components that comprises blade servers, adapters, fabric interconnects, and extenders that are integrated under a common embedded management system. This approach results in far fewer system components and much better manageability, operational efficiencies, and flexibility than comparable data center platforms.
Cisco Unified Computing System Components
Cisco UCS components are shown in Figure 2.
Figure 2 Cisco Unified Computing System Components
The Cisco UCS is designed from the ground up to be programmable and self integrating. A server’s entire hardware stack, ranging from server firmware and settings to network profiles, is configured through model-based management. With Cisco virtual interface cards, even the number and type of I/O interfaces is programmed dynamically, making every server ready to power any workload at any time.
With model-based management, administrators manipulate a model of a desired system configuration, associate a model’s service profile with hardware resources and the system configures itself to match the model. This automation speeds provisioning and workload migration with accurate and rapid scalability. The result is increased IT staff productivity, improved compliance, and reduced risk of failures due to inconsistent configurations.
Cisco Fabric Extender technology reduces the number of system components to purchase, configure, manage, and maintain by condensing three network layers into one. It eliminates both blade server and hypervisor-based switches by connecting fabric interconnect ports directly to individual blade servers and virtual machines. Virtual networks are now managed exactly as physical networks are, but with massive scalability. This represents a radical simplification over traditional systems, reducing capital and operating costs while increasing business agility, simplifying and speeding deployment, and improving performance.
Fabric Interconnect
Cisco UCS Fabric Interconnects creates a unified network fabric throughout the Cisco UCS. They provide uniform access to both networks and storage, eliminating the barriers to deploying a fully virtualized environment based on a flexible, programmable pool of resources.
Cisco Fabric Interconnects comprise a family of line-rate, low-latency, lossless 10-GE, Cisco Data Center Ethernet, and FCoE interconnect switches. Based on the same switching technology as the Cisco Nexus 5000 Series, Cisco UCS 6000 Series Fabric Interconnects provide the additional features and management capabilities that make them the central nervous system of Cisco UCS.
The Cisco UCS Manager software runs inside the Cisco UCS Fabric Interconnects. The Cisco UCS 6000 Series Fabric Interconnects expand the UCS networking portfolio and offer higher capacity, higher port density, and lower power consumption. These interconnects provide the management and communication backbone for the Cisco UCS B-Series Blades and Cisco UCS Blade Server Chassis.
All chassis and all blades that are attached to the Fabric Interconnects are part of a single, highly available management domain. By supporting unified fabric, the Cisco UCS 6200 Series provides the flexibility to support LAN and SAN connectivity for all blades within its domain right at configuration time. Typically deployed in redundant pairs, the Cisco UCS Fabric Interconnect provides uniform access to both networks and storage, facilitating a fully virtualized environment.
The Cisco UCS Fabric Interconnect family is currently comprised of the Cisco 6100 Series and Cisco 6200 Series of Fabric Interconnects.
Cisco UCS 6248UP 48-Port Fabric Interconnect
The Cisco UCS 6248UP 48-Port Fabric Interconnect is a 1 RU, 10-GE, Cisco Data Center Ethernet, FCoE interconnect providing more than 1Tbps throughput with low latency. It has 32 fixed ports of Fibre Channel, 10-GE, Cisco Data Center Ethernet, and FCoE SFP+ ports.
One expansion module slot can be up to sixteen additional ports of Fibre Channel, 10-GE, Cisco Data Center Ethernet, and FCoE SFP+.
Cisco UCS 6248UP 48-Port Fabric Interconnects were used in this study.
Cisco UCS 2200 Series IO Module
The Cisco UCS 2200 Series FEX multiplexes and forwards all traffic from blade servers in a chassis to a parent Cisco UCS Fabric Interconnect over from 10-Gbps unified fabric links. All traffic, even traffic between blades on the same chassis, or VMs on the same blade, is forwarded to the parent interconnect, where network profiles are managed efficiently and effectively by the Fabric Interconnect. At the core of the Cisco UCS Fabric Extender are ASIC processors developed by Cisco that multiplex all traffic.
Note Up to two fabric extenders can be placed in a blade chassis.
UCS 2204 has eight 10GBASE-KR connections to the blade chassis mid-plane, with one connection per fabric extender for each of the chassis’ eight half slots. This gives each half-slot blade server access to each of two 10-Gbps unified fabric-based networks via SFP+ sockets for both throughput and redundancy. It has 4 ports connecting up the fabric interconnect.
UCS 2208 has thirty-two 10GBASE-KR connections to the blade chassis midplane, with one connection per fabric extender for each of the chassis’ eight half slots. This gives each half-slot blade server access to each of two 4x10-Gbps unified fabric-based networks via SFP+ sockets for both throughput and redundancy. It has 8 ports connecting up the fabric interconnect.
Cisco UCS 2208 fabric extenders were utilized in this study.
Cisco UCS Chassis
The Cisco UCS 5108 Series Blade Server Chassis is a 6 RU blade chassis that will accept up to eight half-width Cisco UCS B-Series Blade Servers or up to four full-width Cisco UCS B-Series Blade Servers, or a combination of the two. The UCS 5108 Series Blade Server Chassis can accept four redundant power supplies with automatic load-sharing and failover and two Cisco UCS 2200 series Fabric Extenders. The chassis is managed by Cisco UCS Chassis Management Controllers, which are mounted in the Cisco UCS Fabric Extenders and work in conjunction with the Cisco UCS Manager to control the chassis and its components.
A single UCS managed domain can theoretically scale to up to 40 individual chassis and 320 blade servers. At this time Cisco supports up to 20 individual chassis and 160 blade servers.
Basing the I/O infrastructure on a 10-Gbps unified network fabric allows the Cisco UCS to have a streamlined chassis with a simple yet comprehensive set of I/O options. The result is a chassis that has only five basic components:
- The physical chassis with passive midplane and active environmental monitoring circuitry
- Four power supply bays with power entry in the rear, and hot-swappable power supply units accessible from the front panel
- Eight hot-swappable fan trays, each with two fans
- Two fabric extender slots accessible from the back panel
- Eight blade server slots accessible from the front panel
Cisco UCS B200 M3 Blade Server
Cisco UCS B200 M3 is a third generation half-slot, two-socket Blade Server. The Cisco UCS B200 M3 harnesses the power of the latest Intel® Xeon® processor E5-2600v2 product family, with up to 384 GB of RAM (using 16-GB DIMMs), two optional SAS/SATA/SSD disk drives, and up to dual 4x 10 Gigabit Ethernet throughput, utilizing our VIC 1240 LAN on motherboard (LOM) design. The Cisco UCS B200 M3 further extends the capabilities of Cisco UCS by delivering new levels of manageability, performance, energy efficiency, reliability, security, and I/O bandwidth for enterprise-class virtualization and other mainstream data center workloads.
UCS VIC1240 Converged Network adapter
A Cisco® innovation, the Cisco UCS Virtual Interface Card (VIC) 1240 (Figure 1) is a 4-port 10 Gigabit Ethernet, Fibre Channel over Ethernet (FCoE)-capable modular LAN on motherboard (mLOM) designed exclusively for the M3 generation of Cisco UCS B-Series Blade Servers. When used in combination with an optional Port Expander, the Cisco UCS VIC 1240 capabilities can be expanded to eight ports of 10 Gigabit Ethernet.
The Cisco UCS VIC 1240 enables a policy-based, stateless, agile server infrastructure that can present up to 256 PCIe standards-compliant interfaces to the host that can be dynamically configured as either network interface cards (NICs) or host bus adapters (HBAs). In addition, the Cisco UCS VIC 1240 supports Cisco Data Center Virtual Machine Fabric Extender (VM-FEX) technology, which extends the Cisco UCS fabric interconnect ports to virtual machines, simplifying server virtualization deployment.
Figure 3 Cisco UCS VIC 1240 Converged Network Adapter
Note The UCS VIC1240 virtual interface cards are deployed in the UCS B-Series B200 M3 blade servers.
VMware Horizon View
VMware Horizon View (formerly known as VMware View) simplifies desktop and application management while increasing security and control. Horizon View delivers a personalized high fidelity experience for end-users across sessions and devices. It enables higher availability and agility of desktop services unmatched by traditional PCs while reducing the total cost of desktop ownership up to 50%. End-users can enjoy new levels of productivity and the freedom to access desktops from more devices and locations while giving IT greater policy control.
Horizon View 5.3 Features
Horizon View delivers rich, personalized virtual desktops as a managed service from a virtualization platform built to deliver the entire desktop, including the operating system, applications and data. With VMware Horizon View, desktop administrators virtualize the operating system, applications, and user data and deliver modern desktops to end-users. Get centralized automated management of these components for increased control and cost savings. Improve business agility while providing a flexible high performance desktop experience for end-users, across a variety of network conditions.
Automated Desktop Provisioning
Horizon View Manager provides a single management tool for greater IT efficiency to provision new desktops or groups of desktops, and an easy interface for setting desktop policies. Using a template, you can customize virtual pools of desktops and easily set policies, such as how many virtual machines can be in a pool, or logoff parameters.
Streamlined Application Management
VMware ThinApp application virtualization separates; applications from underlying operating systems and reduces conflict between the OS and other applications for increased compatibility and streamlined management. Applications packaged with ThinApp can be run centrally from the data center, deployed locally to physical or virtual desktops or on USB drives for deployment flexibility.
Advanced Virtual Desktop Image Management
Horizon View Composer enables the rapid creation of desktop images from a golden image. Updates are instant and guaranteed across any number of virtual desktops. When combined with VMware ThinApp , IT administrators can reduce the number of total images, storage requirements and operational costs.
Automate Desktop Operations Management
VMware vCenter Operations Manager for View allows administrators to gain insight into desktop and infrastructure performance, quickly pinpoint and troubleshoot issues. Administrators can optimize resource utilization, and proactively manage the desktop environment through the management dashboards. VMware vCenter Operations Manager for View is an optional add-on for Horizon View customers. You can also leverage PCoIP Extension Services to collect Horizon View statistics into your existing WMI tool.
Efficient Resource Utilization
Horizon View Storage Accelerator optimizes storage load by caching common image blocks when reading virtual desktop images. Space Efficient Disks continuously reduce the storage needed per desktop. Both these technologies improve storage capacity and utilization, thereby reducing costs of additional hardware.
Built-in Security
Maintain control over data and intellectual property by keeping it secure in the data center. Encrypted protocol traffic provides secure end-users access virtual desktops inside or outside of the corporate network. Integration with vShield Endpoint enables offloaded and centralized anti-virus and anti-malware (AV) solutions. To eliminate agent sprawl and AV storm issues, risk of malware infection, and simplify AV administration. Horizon View also supports integration with Radius 2-factor authentication requirements.
What's New in Horizon View 5.3
VMware Horizon View 5.3 delivers important features and enhancements and resolves some known problems in the previous release. TCO was further reduced by optimizing storage reads, improved desktop migration and large scale management, and further enhanced the user-experience with lower bandwidth and client diversity.
This release of VMware Horizon View adds the following new features and support:
- Windows Server 2008 R2 Desktop Operating System Support: Windows Server 2008 R2 (Datacenter edition) is now supported as a desktop operating system. For installation instructions and limitations, see KB 2057605: Using Windows Server 2008 R2 as a desktop operating system in VMware Horizon View .
- Windows 8.1 Desktop Operating System Support: Windows 8.1 is now supported as a desktop operating system.
- VMware Horizon Mirage Support: You can now use VMware Horizon Mirage 4.3 to manage View desktops.
- See the VMware Horizon Mirage Administrator's Guide for complete information about this feature.
- VMware Virtual SAN Datastore Support: When you create a desktop pool in View Administrator; you can now select a Virtual SAN datastore to store desktop virtual machines. Because Virtual SAN is in Beta, this feature is being released as a Tech Preview, which means that it is available for you to try, but it is not recommended for production use and no technical support is provided. The space-efficient virtual disk format is not available on Virtual SAN datastores. If you use Virtual SAN datastore to host virtual desktops, you will not be able to reclaim allocated unused space on the virtual machines.
- View Connection Server Memory Recommendation Messages: If you install View Connection Server with less than 10GB of memory, VMware Horizon View provides memory recommendations by generating warning messages after the installation is complete.
- vDGA Support: vDGA (virtual Dedicated Graphics Acceleration) is now supported for View desktops. For linked-clone desktops, vDGA settings are preserved after refresh, recompose, and rebalance operations. See the VMware white paper Graphics Acceleration in Horizon View Virtual Machines Deployment Guide .
- Linked-Clone Desktop Pool Storage Overcommit Feature Enhancements: The linked-clone desktop pool storage overcommit feature includes a new storage overcommit level called Unbounded. When you select Unbounded, View Manager does not limit the number of linked-clone desktops that it creates based on the physical capacity of the datastore. You select the storage overcommit level for a linked-clone desktop pool on the Select Datastores page when you add or edit a linked-clone pool. Select Unbounded only if you are certain that the datastore has enough storage capacity to accommodate all of the desktops and their future growth.
- View Persona Management Supportability Improvements: Supportability improvements include new log messages and profile size and file and folder count tracking. View Persona Management uses the file and folder counts to suggest folders for redirection in the Windows event log and provides statistics for these folders.
- Support to Grant Domain Administrators Access to Redirected Folders in View Persona Management: A new group policy setting, Add the Administrators group to redirected folders, has been added to make redirected folders accessible to domain administrators. For information about the new group policy setting, see KB 2058932: Granting domain administrators access to redirected folders for View Persona Management .
- VMware Horizon View Agent Direct-Connection Plug-in: You can use VMware Horizon View Agent Direct-Connection Plug-in to connect directly to a virtual desktop. This plug-in is an installable extension to View Agent that allows a View client to directly connect to a View desktop without using View Connection Server. For more information, see VMware Horizon View Agent Direct-Connection Plug-in Administration .
- View Composer Array Integration (VCAI) Support: The Tech Preview designation has been removed from VCAI. VCAI appears as an option during pool creation when you select an NFS datastore on an array that supports VAAI (vStorage API for Array Integration) native snapshots. The VCAI feature is now supported with select NAS vendors. For a list of supported NAS vendors, see KB 2061611: View Composer API for Array Integration (VCAI) support in VMware Horizon View.
- Blast Secure Gateway Maximum Connections: The Blast Secure Gateway (BSG) now supports up to 350 connections to Horizon View desktops from clients using HTML Access. This connection limit applies to a BSG on one View Connection Server instance or security server.
Lower Total Cost of Ownership
Space-efficient disks, native in vSphere, reduce storage costs and administrative overhead by efficiently using and reclaiming storage space to minimize Horizon View Composer image size. This lowers storage capacity requirements for persistent desktops and decreases the need to continuously recompose and restore images
Simplified Management
Improved large-scale management allows customers with large Horizon View deployments to efficiently and logically manage their virtual desktop infrastructure. Overall desktop architecture is simplified with a single VMware vCenter Server™ supporting up to 10,000 desktops in a pod. With support for 32 hosts per pool on VMFS along with NFS and pools spanning multiple VLANs, larger desktop pools can be created to decrease operational costs. Furthermore, View admin UI responsiveness increases and accelerates performance of operations such as provisioning, and rebalance improves the efficiency of the desktop administration team.
VMware vCenter Server virtual appliance support enables greater flexibility in Horizon View infrastructure deployment.
Seamless User Experience
Media services for rich 3D graphics add support for hardware accelerated 3D graphics for the most demanding 3D applications. By virtualizing the graphics processing unit (GPU), you can dedicate or share physical GPU resources across multiple users, providing a rich 3D experience from the data center. Using a combination of software and hardware-accelerated graphics, VMware Horizon View™ provides the greatest flexibility for delivering 3D graphics for virtual desktops and workstation use cases. 3D graphics acceleration is built upon the VMware vSphere® platform. Only Horizon View is designed to fully leverage vSphere, expanding the value of combined solutions.
Horizon View media services already support unified communications for Cisco Unified Communications.
Horizon View HTML access enables users to access desktops based on Horizon View from HTML5-capable browsers to securely access their data and applications. Without requiring the installation of any software or plug-ins, end users conveniently can access their desktops on any device. With Horizon Workspace™ integration, that same desktop convenience is expanded to provide end users access to their desktops, data and apps, all from a single location. Horizon View HTML access is available in the Horizon View feature pack.
VMware Horizon View Clients for iOS and Android with Unity make it easier than ever to access Windows applications on your iPhone, iPad or Android device. Remove the frustration of working with Windows on mobile devices with a new mobile native user interface. With Unity users can easily browse, search, and open Windows applications and files, set applications and files as favorites, and easily switch between running applications.
Windows 8 support gives users ability to use the latest OS inside their virtual desktops. Horizon View Client has also been updated to run on the latest Windows 8 devices.
VMware Horizon View 5.3 Hosted Virtual Desktop (HVD) Overview
Hosted Virtual Desktop (HVD) uses a hypervisor to host all the desktops in the data center.
Three types of HVD pools are available with Horizon View 5.3: Automated, Manual, and Terminal Services Pools. These pool types are discussed below.
- Automated HVD pools use Horizon View Composer to create some number of HVDs. HVD users can be assigned as floating or dedicated users. Floating users will be assigned randomly to HVDs as they log on. Once the user logs off, the HVD is available for any other user. Dedicated user assignments insure that a user is provided the same HVD each time he or she connects to the Horizon View Connection server. Automated pools can utilize the PCoIP protocol and View Persona Management.
- Automated HVD pools can create two types of HVDs: Full virtual machines created from a vCenter template or Horizon View Composer linked clones which share the same base image and use less storage.
- Manual HVD pools provide access to an existing set of HVDs. Any type of machine that can install the Horizon View Agent is supported. Examples could include vCenter virtual machines, physical machines, or blade PCs. Manual pools support the PCoIP protocol, View Persona Management, and Local Mode.
- Microsoft Terminal Services Pools provide Terminal Services sessions as desktops to Horizon View users. The Horizon View Connection Server manages these sessions in the same way it does for Automated or Manual HVD pools. Terminal Services Pools support View Persona Management.
For this VSPEX study, we utilized Automated HVD pools with floating user assignments and Horizon View Composer linked clones over the PCoIP protocol.
View Persona Manager was not deployed.
The following figure shows the logical architecture for a Horizon View 5.3 deployment, including the optional related product; Thin App. Thin App provides application streaming capability and is not included in this study.
Figure 4 VMware Horizon View 5 Architecture Diagram
EMC VNX Series
The VNX series delivers uncompromising scalability and flexibility for the mid-tier while providing market-leading simplicity and efficiency to minimize total cost of ownership. Customers can benefit from VNX features such as:
- Next-generation unified storage, optimized for virtualized applications.
- Extended cache by using Flash drives with Fully Automated Storage Tiering for Virtual Pools (FAST VP) and FAST Cache that can be optimized for the highest system performance and lowest storage cost simultaneously on both block and file.
- Multiprotocol supports for file, block, and object with object access through EMC Atmos™ Virtual Edition (Atmos VE).
- Simplified management with EMC Unisphere™ for a single management framework for all NAS, SAN, and replication needs.
- Up to three times improvement in performance with the latest Intel Xeon multicore processor technology, optimized for Flash.
- 6 Gb/s SAS back end with the latest drive technologies supported:
- 3.5” 100 GB and 200 GB Flash, 3.5” 300 GB, and 600 GB 15k or 10k rpm SAS, and 3.5” 1 TB, 2 TB and 3 TB 7.2k rpm NL-SAS
- 2.5” 100 GB and 200 GB Flash, 300 GB, 600 GB and 900 GB 10k rpm SAS
- Expanded EMC UltraFlex™ I/O connectivity—Fibre Channel (FC), Internet Small Computer System Interface (iSCSI), Common Internet File System (CIFS), network file system (NFS) including parallel NFS (pNFS), Multi-Path File System (MPFS), and Fibre Channel over Ethernet (FCoE) connectivity for converged networking over Ethernet.
The VNX series includes five software suites and three software packs that make it easier and simpler to attain the maximum overall benefits.
–
VNX FAST Suite—Automatically optimizes for the highest system performance and the lowest storage cost simultaneously (FAST VP is not part of the FAST Suite for VNX5100™).
–
–
–
–
–
–
–
EMC VNX5600 Used in Testing
EMC VNX 5600 is a unified storage platform for multi-protocol file, block and object storage. It is powered by Intel quad-core Xeon 5600 series processors and delivers five 9’s availability. It is designed to deliver maximum performance and scalability for enterprise and mid-tier companies, enabling them to dramatically grow, share, and cost-effectively manage multi-protocol file and block systems. It supports up to 250 drives and three X-Blades (also known as Data Movers) for file protocol support. This solution was validated Fibre Channel for hypervisor SAN boot, data storage of virtual desktops, SQL database, and infrastructure virtual machines such Horizon View Connection Servers, Horizon View Composer Servers, VMware vCenter Servers, and other supporting services. An NFS or iSCSI variant could be deployed on the VNX5600 using NFS or iSCSI for data storage of virtual desktops.
VMware ESXi 5.5
VMware, Inc. provides virtualization software. VMware’s enterprise software hypervisors for servers—VMware vSphere ESX, VMware vSphere ESXi, and VSphere—are bare-metal hypervisors that run directly on server hardware without requiring an additional underlying operating system.
VMware on ESXi 5.5 Hypervisor
ESXi 5.5 is a "bare-metal" hypervisor, so it installs directly on top of the physical server and partitions it into multiple virtual machines that can run simultaneously, sharing the physical resources of the underlying server. VMware introduced ESXi in 2007 to deliver industry-leading performance and scalability while setting a new bar for reliability, security and hypervisor management efficiency.
Due to its ultra-thin architecture with less than 100MB of code-base disk footprint, ESXi delivers industry-leading performance and scalability plus:
- Improved Reliability and Security — with fewer lines of code and independence from general purpose OS, ESXi drastically reduces the risk of bugs or security vulnerabilities and makes it easier to secure your hypervisor layer.
- Streamlined Deployment and Configuration — ESXi has far fewer configuration items than ESX, greatly simplifying deployment and configuration and making it easier to maintain consistency.
- Higher Management Efficiency — The API-based, partner integration model of ESXi eliminates the need to install and manage third party management agents. You can automate routine tasks by leveraging remote command line scripting environments such as vCLI or PowerCLI.
- Simplified Hypervisor Patching and Updating — Due to its smaller size and fewer components, ESXi requires far fewer patches than ESX, shortening service windows and reducing security vulnerabilities.
Modular Architecture
Today’s IT departments are facing a rapidly-evolving workplace environment. The workforce is becoming increasingly diverse and geographically distributed and includes offshore contractors, distributed call center operations, knowledge and task workers, partners, consultants, and executives connecting from locations around the globe at all times.
An increasingly mobile workforce wants to use a growing array of client computing and mobile devices that they can choose based on personal preference. These trends are increasing pressure on IT to ensure protection of corporate data and to prevent data leakage or loss through any combination of user, endpoint device, and desktop access scenarios (Figure xx-4.5.1). These challenges are compounded by desktop refresh cycles to accommodate aging PCs and bounded local storage and migration to new operating systems, specifically Microsoft Windows 7.
Figure 5 The Evolving Workplace Landscape
Some of the key drivers for desktop virtualization are increased data security and reduced TCO through increased control and reduced management costs.
Cisco Data Center Infrastructure for Desktop Virtualization
Cisco focuses on three key elements to deliver the best desktop virtualization data center infrastructure: simplification, security, and scalability. The software combined with platform modularity provides a simplified, secure, and scalable desktop virtualization platform (Figure xx-4.5.1.1).
Figure 6 VMware Horizon View 5.3 on Cisco UCS
Simplified
Cisco UCS provides a radical new approach to industry standard computing and provides the heart of the data center infrastructure for desktop virtualization and the Cisco Virtualization Experience (VXI). Among the many features and benefits of Cisco UCS are the drastic reductions in the number of servers needed and number of cables per server and the ability to very quickly deploy or re-provision servers through Cisco UCS Service Profiles. With fewer servers and cables to manage and with streamlined server and virtual desktop provisioning, operations are significantly simplified. Thousands of desktops can be provisioned in minutes with Cisco Service Profiles and Cisco storage partners’ storage-based cloning. This speeds time to productivity for end users, improves business agility, and allows IT resources to be allocated to other tasks.
IT tasks are further simplified through reduced management complexity, provided by the highly integrated Cisco UCS Manager, along with fewer servers, interfaces, and cables to manage and maintain. This is possible due to the industry-leading, highest virtual desktop density per blade of Cisco UCS along with the reduced cabling and port count due to the unified fabric and unified ports of Cisco UCS and desktop virtualization data center infrastructure.
Simplification also leads to improved and more rapid success of a desktop virtualization implementation. Cisco and its partners –VMware and EMC – have developed integrated, validated architectures, including available pre-defined, validated infrastructure packages, known as Cisco Solutions for EMC VSPEX End User Computing.
Secure
While virtual desktops are inherently more secure than their physical world predecessors, they introduce new security considerations. Desktop virtualization significantly increases the need for virtual machine-level awareness of policy and security, especially given the dynamic and fluid nature of virtual machine mobility across an extended computing infrastructure. The ease with which new virtual desktops can proliferate magnifies the importance of a virtualization-aware network and security infrastructure. Cisco UCS and Nexus data center infrastructure for desktop virtualization provides stronger data center, network, and desktop security with comprehensive security from the desktop to the hypervisor. Security is enhanced with segmentation of virtual desktops, virtual machine-aware policies and administration, and network security across the LAN and WAN infrastructure.
Scalable
Growth of a desktop virtualization solution is all but inevitable and it is critical to have a solution that can scale predictably with that growth. The Cisco solution supports more virtual desktops per server and additional servers scale with near linear performance. Cisco data center infrastructure provides a flexible platform for growth and improves business agility. Cisco UCS Service Profiles allow for on-demand desktop provisioning, making it easy to deploy dozens or thousands of additional desktops.
Each additional Cisco UCS server provides near linear performance and utilizes Cisco’s dense memory servers and unified fabric to avoid desktop virtualization bottlenecks. The high performance, low latency network supports high volumes of virtual desktop traffic, including high resolution video and communications.
Cisco UCS and Nexus data center infrastructure is an ideal platform for growth, with transparent scaling of server, network, and storage resources to support desktop virtualization.
Savings and Success
As demonstrated above, the simplified, secure, scalable Cisco data center infrastructure solution for desktop virtualization will save time and cost. There will be faster payback, better ROI, and lower TCO with the industry’s highest virtual desktop density per server resulting in fewer servers required, reducing both capital expenditures (CapEx) and operating expenditures (OpEx). There will also be much lower network infrastructure costs, with fewer cables per server and fewer ports required, via the Cisco UCS architecture and unified fabric.
The simplified deployment of Cisco UCS for desktop virtualization speeds up time to productivity and enhances business agility. IT staff and end users are more productive more quickly and the business can react to new opportunities by simply deploying virtual desktops whenever and wherever they are needed. The high performance Cisco systems and network deliver a near-native end-user experience, allowing users to be productive anytime, anywhere.
Understanding Desktop User Groups
There must be a considerable effort within the enterprise to identify desktop user groups and their memberships. The most broadly recognized, high level user groups are:
- Task Workers Groups of users working in highly specialized environments where the number of tasks performed by each worker is essentially identical. These users are typically located at a corporate facility (e.g., call center employees).
- Knowledge/Office Workers Groups of users who use a relatively diverse set of applications that are Web-based and installed and whose data is regularly accessed. They typically have several applications running simultaneously throughout their workday and a requirement to utilize Flash video for business purposes. This is not a singular group within an organization. These workers are typically located at a corporate office (e.g., workers in accounting groups).
- Power Users Groups of users who run high-end, memory, processor, disk IO, and/or graphic-intensive applications, often simultaneously. These users have high requirements for reliability, speed, and real-time data access (e.g., design engineers).
- Mobile Workers Groups of users who may share common traits with Knowledge/Office Workers, with the added complexity of needing to access applications and data from wherever they are whether at a remote corporate facility, customer location, at the airport, at a coffee shop, or at home all in the same day (e.g., a company’s outbound sales force).
- Remote Workers Groups of users who could fall into the Task Worker or Knowledge/Office Worker groups but whose experience is from a remote site that is not corporate owned, most often from the user’s home. This scenario introduces several challenges in terms of type, available bandwidth, and latency and reliability of the user’s connectivity to the data center (e.g., a work-from-home accounts payable representative).
- Guest/Contract Workers Groups of users who need access to a limited number of carefully controlled enterprise applications and data and resources for short periods of time. These workers may need access from the corporate LAN or remote access (for example, a medical data transcriptionist).
There is good reason to search for and identify multiple sub-groups of the major groups listed above in the enterprise. Typically, each sub-group has different application and data requirements.
Understanding Applications and Data
Once the desktop user groups and sub-groups have been identified, the next task is to catalog group application and data requirements. This can be one of the most time-consuming processes in the VDI planning exercise, but is essential for the VDI project’s success. If the applications and data are not identified and co-located, performance will be negatively affected.
The process of analyzing the variety of application and data pairs for an organization will likely be complicated by the inclusion cloud applications, like SalesForce.com. This application and data analysis is beyond the scope of this Cisco Validated Design, but should not be omitted from the planning process. There are a variety of third party tools available to assist organizations with this crucial exercise.
Project Planning and Solution Sizing Sample Questions
Now that user groups, their applications and their data requirements are understood, some key project and solution sizing questions may be considered.
General project questions should be addressed at the outset, including:
- Has a VDI pilot plan been created based on the business analysis of the desktop groups, applications and data?
- Is there infrastructure and budget in place to run the pilot program?
- Are the required skill sets to execute the VDI project available? Can we hire or contract for them?
- Do we have end user experience performance metrics identified for each desktop sub-group?
- How will we measure success or failure?
- What is the future implication of success or failure?
Provided below is a short, non-exhaustive list of sizing questions that should be addressed for each user sub-group:
- What is the desktop OS planned? Windows 7 or Windows XP?
- 32 bit or 64 bit desktop OS?
- How many virtual desktops will be deployed in the pilot? In production? All Windows 7?
- How much memory per target desktop group desktop?
- Are there any rich media, Flash, or graphics-intensive workloads?
- What is the end point graphics processing capability?
- Will Thin App be used for streamed applications or will all applications be installed in the image?
- Will you use floating assignment or assigned user desktops?
- Will you use linked clone or full copy desktops?
- How will you manage user persona?
- What is the storage configuration in the existing environment?
- Are there sufficient IOPS available for the write-intensive VDI workload?
- Will there be storage dedicated and tuned for VDI service?
- Is there a voice component to the desktop?
- Is anti-virus a part of the image?
- Is user profile management (e.g., non-roaming profile based) part of the solution?
- What is the fault tolerance, failover, disaster recovery plan?
- Are there additional desktop sub-group specific questions?
Proof of Concept/Pilot
To validate what you have learned during your analysis, create an isolated Proof of Concept environment to test the various workloads and validate your sizing calculations.
Then create a Pilot environment and get users from each user group who can exercise all of the organizations key applications. Use the pilot user feedback to further refine you design in preparation for production roll out.
Failure to follow these key steps will make a successful virtual desktop deployment project nearly impossible.
Cisco Services
Cisco offers assistance for customers in the analysis, planning, implementation, and support phases of the VDI lifecycle. These services are provided by the Cisco Advanced Services group. Some examples of Cisco services include:
The Solution: A Unified, Pre-Tested and Validated Infrastructure
To meet the challenges of designing and implementing a modular desktop infrastructure, Cisco, EMC and VMware have collaborated to create the data center solution for virtual desktops outlined in this document.
Cisco Networking Infrastructure
This section describes the Cisco networking infrastructure components used in the configuration.
Cisco Nexus 5548 Switch
The Cisco Nexus 5548UP is a 1-RU 10 Gigabit Ethernet, Fibre Channel, and FCoE switch offering up to 960 Gbps of throughput and up to 48 ports. The switch has 32 unified ports that accept modules and cables meeting the Small Form-Factor Pluggable Plus (SFP+) standard and one expansion slot. Expansion slot options include:
- Ethernet module that provides sixteen 1/10 Gigabit Ethernet and FCoE ports using the SFP+ interface.
- Fibre Channel plus Ethernet module that provides eight 1/10 Gigabit Ethernet and FCoE ports using the SFP+ interface, and eight ports of 8/4/2/1-Gbps native Fibre Channel connectivity using the SFP+/SFP interface.
- Unified port module that provides up to sixteen 1/10 Gigabit Ethernet and FCoE ports using the SFP+ interface or up to sixteen ports of 8/4/2/1-Gbps native Fibre Channel connectivity using the SFP+ and SFP interfaces; the use of 1/10 Gigabit Ethernet or 8/4/2/1-Gbps Fibre Channel on a port is mutually exclusive but can be selected for any of the 16 physical ports per module.
- Four port QSFP Ethernet module that provides 4 40 Gigabit Ethernet ports using QSFP interface.
The switch has a single serial console port and a single out-of-band 10/100/1000-Mbps Ethernet management port. Two N+1 redundant, hot-pluggable power supplies and five N+1 redundant, hot-pluggable fan modules provide highly reliable front-to-back cooling.
Cisco Nexus 5500 Series Feature Highlights
The switch family's rich feature set makes the series ideal for rack-level, access-layer applications. It protects investments in data center racks with standards-based Ethernet and FCoE features that allow IT departments to consolidate networks based on their own requirements and timing.
- The combination of high port density, wire-speed performance, and extremely low latency makes the switch an ideal product to meet the growing demand for 10 Gigabit Ethernet at the rack level. The switch family has sufficient port density to support single or multiple racks fully populated with blade and rack-mount servers.
- Built for today‘s data centers, the switches are designed just like the servers they support. Ports and power connections are at the rear, closer to server ports, helping keep cable lengths as short and efficient as possible. Hot-swappable power and cooling modules can be accessed from the front panel, where status lights offer an at-a-glance view of switch operation. Front-to-back cooling is consistent with server designs, supporting efficient data center hot-aisle and cold-aisle designs. Serviceability is enhanced with all customer replaceable units accessible from the front panel. The use of SFP+ ports offers increased flexibility to use a range of interconnect solutions, including copper for short runs and fibre for long runs.
- FCoE and IEEE data center bridging features support I/O consolidation, ease management of multiple traffic flows, and optimize performance. Although implementing SAN consolidation requires only the lossless fabric provided by the Ethernet pause mechanism, the Cisco Nexus 5500 Series switches provide additional features that create an even more easily managed, high-performance, unified network fabric.
Specific features and benefits provided by the Cisco Nexus 5500 Series follow.
10GB Ethernet, FCoE, and Unified Fabric Features
The switch series, using cut-through architecture, supports line-rate 10 Gigabit Ethernet on all ports while maintaining consistently low latency independent of packet size and services enabled. It supports a set of network technologies known collectively as Data Center Bridging (DCB) that increases the reliability, efficiency, and scalability of Ethernet networks. These features allow the switches to support multiple traffic classes over a lossless Ethernet fabric, thus enabling consolidation of LAN, SAN, and cluster environments. Its ability to connect Fibre Channel over Ethernet (FCoE) to native Fibre Channel protects existing storage system investments while dramatically simplifying in-rack cabling.
Low Latency
The cut-through switching technology used in the Cisco Nexus 5500 Series ASICs enables the product to offer a low latency of 3.2 microseconds, which remains constant regardless of the size of the packet being switched. This latency was measured on fully configured interfaces, with access control lists (ACLs), QoS, and all other data path features turned on. The low latency on the Cisco Nexus 5500 Series enables application-to-application latency on the order of 10 microseconds (depending on the NIC). These numbers, together with the congestion management features described in the next section, make the Cisco Nexus 5500 Series a great choice for latency-sensitive environments.
Cisco Nexus 1000V Feature Highlights
Cisco Nexus 1000V Series Switches are virtual machine access switches that are an intelligent software switch implementation based on IEEE 802.1Q standard for VMware vSphere environments running the Cisco® NX-OS Software operating system. Operating inside the VMware ESX hypervisor, the Cisco Nexus 1000V Series supports Cisco VN-Link server virtualization technology to provide:
- Policy-based virtual machine connectivity
- Mobile virtual machine security and network policy
- Non-disruptive operational model for server virtualization and networking teams
With the Cisco Nexus 1000V Series, you can have a consistent networking feature set and provisioning process all the way from the virtual machine access layer to the core of the data center network infrastructure. Virtual servers can now use the same network configuration, security policy, diagnostic tools, and operational models as their physical server counterparts attached to dedicated physical network ports. Virtualization administrators can access pre-defined network policy that follows mobile virtual machines to help ensure proper connectivity, saving valuable time for virtual machine administration.
Developed in close collaboration with VMware, the Cisco Nexus 1000V Series is certified by VMware to be compatible with VMware vSphere, vCenter, ESX, and ESXi, and with many other vSphere features. You can use the Cisco Nexus 1000V Series to manage your virtual machine connectivity with confidence in the integrity of the server virtualization infrastructure.
The Cisco Nexus 1000V Release 4.2(1)SV2(1.1) software onwards, is being offered in two editions:
- Cisco Nexus 1000V Essential Edition: This is available at no cost and provides most of the comprehensive Layer 2 networking features of the Cisco Nexus 1000V Series, including VXLAN, Cisco vPath for service insertion and chaining, and VMware vCloud Director integration.
- Cisco Nexus 1000V Advanced Edition: This version offers value-added security features such as Domain Host Control Protocol (DHCP) snooping, IP source guard, Dynamic Address Resolution Protocol (ARP) Inspection, and Cisco TrustSec® Secure Group Access (SGA) support (a new feature in Release 2.1). The Cisco VSG zone-based virtual firewall is also included in the Advanced Edition.
Nexus 1000V Product Architecture
Cisco Nexus 1000V Series Switches have two major components: the Virtual Ethernet Module (VEM), which runs inside the hypervisor, and the external Virtual Supervisor Module (VSM), which manages the VEMs.
The Cisco Nexus 1000V Series VEM runs as part of the VMware ESX or ESXi kernel and replaces the VMware virtual switch (vSwitch). This level of integration helps ensure that the Cisco Nexus 1000V Series is fully aware of all server virtualization events, such as VMware vMotion and Distributed Resource Scheduler (DRS). The VEM takes configuration information from the VSM and provides advanced networking functions: quality of service (QoS), security features, and monitoring features.
Virtual Supervisor Module (VSM)
The Cisco Nexus 1000V Series VSM controls multiple VEMs as one logical modular switch. Configuration is performed through the VSM and is automatically propagated to the VEMs. Instead of configuring soft switches inside the hypervisor on a host-by-host basis administrators can define configurations for immediate use on all VEMs being managed by the VSM from a single interface.
Nexus 1000V Features and Benefits
The Cisco Nexus 1000V Series provides a common management model for both physical and virtual network infrastructures through Cisco VN-Link technology, which includes policy-based virtual machine connectivity, mobility of virtual machine security and network properties, and a non-disruptive operational model.
Policy-Based Virtual Machine Connectivity
To facilitate easy creation and provisioning of virtual machines, the Cisco Nexus 1000V Series includes port profiles. Port profiles enable you to define virtual machine network policies for different types or classes of virtual machines and then apply the profiles through the VMware vCenter. Port profiles are a scalable mechanism for configuring networks with large numbers of virtual machines. When the Port Profiles include QoS and security policies, they formulate a complete service-level agreement (SLA) for the virtual machine’s traffic.
Mobility of Virtual Machine Security and Network Properties
Network and security policies defined in the port profile follow the virtual machine throughout its lifecycle, whether it is being migrated from one server to another, suspended, hibernated, or restarted. In addition to migrating the policy, the Cisco Nexus 1000V Series VSM moves the virtual machine’s network state. Virtual machines participating in traffic-monitoring activities can continue these activities uninterrupted by VMware vMotion operations. When a specific port profile is updated, the Cisco Nexus 1000V Series automatically provides live updates to all the virtual ports using that same port profile. The capability to migrate network and security policies through VMware vMotion makes regulatory compliance much easier to enforce with the Cisco Nexus 1000V Series because the security policy is defined in the same way as for physical servers and is constantly enforced by the switch.
Besides traditional switching capability, the Cisco Nexus 1000V Series offers the Cisco vPath architecture to support virtualized network services with:
- Intelligent Traffic Steering: This feature redirects packets in a network flow to a virtual service virtual machine called a Virtual Service Node (VSN), which can be on a different server. Thus, a VSN is not required on every server, providing flexible and consolidated deployment.
- Performance Acceleration: VEM caches the VSN’s decision for a flow, implements the service in all subsequent packets of the flow, and accelerates virtualized network service in the hypervisor kernel.
Cisco Virtual Service Gateway (VSG) is the first VSN to leverage the Cisco vPath architecture and provides multi-tenant, scalable, security services for virtual machines on the Cisco Nexus 1000V Series Switches.
Non-disruptive Operational Model
Because of its close integration with VMware vCenter, the Cisco Nexus 1000V Series allows virtualization administrators to continue using VMware tools to provision virtual machines. At the same time, network administrators can provision and operate the virtual machine network the same way they do the physical network. While both teams work independently, the Cisco Nexus 1000V Series enforces consistent configuration and policy throughout the server virtualization environment. This level of integration lowers the cost of ownership while supporting organizational boundaries among server, network, security, and storage teams.
Inside VMware vCenter, virtual machines are configured as before. For network configuration, port profiles defined on the Cisco Nexus 1000V Series VSM are displayed by VMware vCenter as port groups. Virtualization administrators can take advantage of preconfigured port groups and focus on virtual machine management, and network administrators can use port profiles to apply policy for a large number of ports at the same time. Together, both teams can deploy server virtualization more efficiently and with lower operating costs.
- Optimized server bandwidth for I/O-intensive applications: Today, network interfaces are often dedicated to a particular type of traffic, such as VMware Console or vMotion. With the Cisco Nexus 1000V Series, all network interface cards (NICs) can be treated as a single logical channel with QoS attached to each type of traffic. Consequently, the bandwidth to the server can be more efficiently utilized, with network-intensive applications virtualized.
- Easier security audits with consistent security policy: Security audits on virtual machines are usually more difficult to perform because virtual machines are secured differently than physical servers. As the Cisco Nexus 1000V Series provides persistent security policy to mobile virtual machines, security audits are similar to those for physical servers.
- Virtual machine as basic building block of data center: With the Cisco Nexus 1000V Series, virtual machines are treated the same way as physical servers in security policy, monitoring and troubleshooting, and the operational model between network and server administrators, enabling virtual machines to be true basic building blocks of the data center. These operational efficiencies lead to greater scaling of server virtualization deployments with lower operating expenses.
The Cisco Nexus 1000V Series is compatible with VMware vSphere as a VMware vNetwork Distributed Switch (vDS) with support for VMware ESX and ESXi hypervisors and integration with VMware vCenter Server. Cisco Nexus 1000V Series Switches are compatible with the various VMware vSphere features.
Architecture and Design of Horizon View 5.3 on Cisco Unified Computing System and EMC VNX Storage
Design Fundamentals
There are many reasons to consider a virtual desktop solution such as an ever growing and diverse base of user devices, complexity in management of traditional desktops, security, and even Bring Your Own Computer (BYOC) to work programs. The first step in designing a virtual desktop solution is to understand the user community and the type of tasks that are required to successfully execute their role. The following user classification is provided:
Knowledge Workers today do not just work in their offices all day – they attend meetings, visit branch offices, work from home, and even coffee shops. These anywhere workers expect access to all of their same applications and data wherever they are.
External Contractors are increasingly part of your everyday business. They need access to certain portions of your applications and data, yet administrators still have little control over the devices they use and the locations they work from. Consequently, IT is stuck making trade-offs on the cost of providing these workers a device vs. the security risk of allowing them access from their own devices.
Task Workers perform a set of well-defined tasks. These workers access a small set of applications and have limited requirements from their PCs. However, since these workers are interacting with your customers, partners, and employees, they have access to your most critical data.
Mobile Workers need access to their virtual desktop from everywhere, regardless of their ability to connect to a network. In addition, these workers expect the ability to personalize their PCs, by installing their own applications and storing their own data, such as photos and music, on these devices.
Shared Workstation users are often found in state-of-the-art University and business computer labs, conference rooms or training centers. Shared workstation environments have the constant requirement to re-provision desktops with the latest operating systems and applications as the needs of the organization change, tops the list.
After the user classifications have been identified and the business requirements for each user classification have been defined, it becomes essential to evaluate the types of virtual desktops that are needed based on user requirements. There are essentially five potential desktops environments for each user:
- Traditional PC: A traditional PC is what ¯typically constituted a desktop environment: physical device with a locally installed operating system.
- Hosted Shared Desktop: A hosted, server-based desktop is a desktop where the user interacts through a delivery protocol. With hosted, server-based desktops, a single installed instance of a server operating system, such as Microsoft Windows Server 2008 R2, is shared by multiple users simultaneously. Each user receives a desktop “session” and works in an isolated memory space. Changes made by one user could impact the other users.
- Hosted Virtual Desktop: A hosted virtual desktop is a virtual desktop running either on virtualization layer (ESX) or on bare metal hardware. The user does not work with and sit in front of the desktop, but instead the user interacts through a delivery protocol.
- Streamed Applications: Streamed desktops and applications run entirely on the user‘s local client device and are sent from a server on demand. The user interacts with the application or desktop directly but the resources may only available while they are connected to the network.
- Local Virtual Desktop: A local virtual desktop is a desktop running entirely on the user‘s local device and continues to operate when disconnected from the network. In this case, the user’s local device is used as a type 1 hypervisor and is synced with the data center when the device is connected to the network.
For the purposes of the validation represented in this document only hosted virtual desktops were validated. Each of the sections provides some fundamental design decisions for this environment.
Hosted Virtual Desktop (HVD) Design Fundamentals
VMware Horizon View 5.3 can be used to deliver a variety of virtual desktop configurations. When evaluating a HVD deployment, consider the following:
Choosing a Display Protocol
A display protocol provides end users with a graphical interface to a View desktop that resides in the data center. You can use PCoIP (PC-over-IP), which VMware provides, or Microsoft RDP (Remote Desktop Protocol.) You can set policies to control which protocol is used or to allow end users to choose the protocol when they login to a desktop.
Note For this study, we used the PCoIP protocol.
VMware View with PCoIP
PCoIP provides an optimized desktop experience for the delivery of the entire desktop environment, including applications, images, audio, and video content for a wide range of users on the LAN or across the WAN. PCoIP can compensate for an increase in latency or a reduction in bandwidth, to ensure that end users can remain productive regardless of network conditions.
PCoIP is supported as the display protocol for View desktops with virtual machines and with physical machines that contain Teradici host cards.
Key features of PCoIP include the following:
- For users outside the corporate firewall, you can use this protocol with your company's virtual private network or with View security servers.
- Advanced Encryption Standard (AES) 128-bit encryption is supported and is turned on by default.
- Connections from all types of View clients. For more information, go to: https://www.vmware.com/support/viewclients/doc/viewclients_pubs.html
- USB redirection is supported.
- Audio redirection with dynamic audio quality adjustment for LAN and WAN is supported.
- Optimization controls for reducing bandwidth usage on the LAN and WAN.
- Multiple monitors are supported. You can use up to four monitors and adjust the resolution for each monitor separately, with a resolution of up to 2560x1600 per display. Pivot display and autofit are also supported. When the 3D feature is enabled, up to 2 monitors are supported with a resolution of up to 1920x1200.
- 32-bit color is supported for virtual displays.
- Clear Type fonts are supported.
- Copy and paste of text and images between a local Windows client system and the desktop is supported, up to 1MB. Supported file formats include text, images, and RTF (Rich Text Format). You cannot copy and paste system objects such as folders and files between systems.
- 480p-formatted video: You can play video at 480p or lower at native resolutions when the View desktop has a single virtual CPU. If the operating system is Windows 7 and you want to play the video in high-definition Flash or in full screen mode, the desktop requires a dual virtual CPU.
- 720p-formatted video: You can play video at 720p at native resolutions if the View desktop has a dual virtual CPU. Performance might be affected if you play videos at 720p in high definition or in full screen mode.
- 1080p-formatted video: If the View desktop has a dual virtual CPU, you can play 1080p formatted video, although the media player might need to be adjusted to a smaller window size.
- 3D: If you plan to use 3D applications such as Windows Aero themes or Google Earth, the Windows 7 View desktop must have virtual hardware version 8, available with vSphere 5 and later. You must also turn on the pool setting called Windows 7 3D Rendering. Up to 2 monitors are supported, and the maximum screen resolution is 1920 x 1200. This non-hardware accelerated graphics feature enables you to run DirectX 9 and OpenGL 2.1 applications without requiring a physical graphics processing unit (GPU).
Recommended guest operating system settings include the following settings:
Microsoft RDP
Remote Desktop Protocol is the same multichannel protocol many people already use to access their work computer from their home computer. Microsoft Remote Desktop Connection (RDC) uses RDP to transmit data.
Microsoft RDP provides the following features:
- With RDP 6, you can use multiple monitors in span mode. RDP 7 has true multiple monitor support, for up to 16 monitors.
- You can copy and paste text and system objects such as folders and files between the local system and the View desktop.
- RDP supports 32-bit color.
- RDP supports 128-bit encryption.
- You can use this protocol for making secure, encrypted connections to a View security server in the corporate DMZ.
Following are RDP-related requirements and considerations for different Windows operating systems and features:
- For Windows XP and Windows XP Embedded systems, you should use Microsoft RDC 6.x.
- Windows Vista comes with RDC 6.x installed, though RDC 7 is recommended.
- Windows 7 comes with RDC 7 installed. Windows 7 SP1 comes with RDC 7.1 installed.
- You must have RDC 6.0 or later to use multiple monitors.
- For Windows XP desktop virtual machines, you must install the RDP patches listed in Microsoft Knowledge Base (KB) articles 323497 and 884020. If you do not install the RDP patches, a Windows Sockets failed error message might appear on the client.
- The View Agent installer configures the local firewall rule for inbound RDP connections to match the current RDP port of the host operating system, which is typically 3389. If you change the RDP port number, you must change the associated firewall rules.
Note You can download RDC versions from the Microsoft Web site.
Choose a User Profile Management System
There are a number of options for managing user profiles for HVDs. The two methods we considered for this study were Microsoft Roaming User Profiles and View Persona Manager. It is important to select and deploy a method so that user settings for software applications and user preferences are maintained, particularly for floating desktops. Both methods are discussed briefly below. (We used Microsoft Roaming User Profiles in the study.)
Microsoft Roaming User Profiles and Folder Redirection
This technology has been around for more than a dozen years. It was significantly enhanced with the introduction of Windows Vista and updated again with Windows 7. Version two (v2) roaming profiles were introduced, adding 8 additional folders that can be redirected. This greatly reduces the time it takes to load the user’s profile during logon.
Using Roaming User Profiles and Folder redirection require a network shares that all users have access to during the virtual desktop session. The user must have read and write access to their profile folder and folder redirection folder, which get created on first login after Roaming User Profiles is configured.
Utilizing Microsoft Active Directory Group Policy is the recommended method for providing Roaming User Profiles and Folder Redirection to your users. See the article titled Managing Roaming User Data Deployment Guide at the following URL for details on how to configure both Roaming User Profiles and Folder Redirection:
http://technet.microsoft.com/en-us/library/cc766489%28WS.10%29.aspx
Note Even though the article talks about Windows Vista, the product where the significant changes to Roaming User Profiles and Folder Redirection were first implemented, it applies to Windows 7 as well.
VMware Persona Management
You can use View Persona Management with View desktops and with physical computers and virtual machines that are not managed by View. View Persona Management retains changes that users make to their profiles. User profiles comprise a variety of user-generated information.
- User-specific data and desktop settings, which allow the desktop appearance to be the same regard less of which desktop a user logs in to.
- Application data and settings. For example, these settings allow applications to remember toolbar positions and preferences.
- Windows registry entries configured by user applications.
To facilitate these abilities, View Persona Management requires storage on a CIFS share equal or greater than the size of the user's local profile.
Minimizing Logon and Logoff Times
View Persona Management minimizes the time it takes to log on to and off of desktops.
- View takes recent changes in the profile on the View desktop and copies them to the remote repository at regular intervals. The default is every 10 minutes. In contrast, Windows roaming profiles wait until logoff time and copy all changes to the server at logoff.
- During logon, View downloads only the files that Windows requires, such as user registry files. Other files are copied to the View desktop when the user or an application opens them from the profile folder in the View desktop.
- With View Persona Management, during logoff, only files that were updated since the last replication are copied to the remote repository.
With View Persona Management, you can avoid making any changes to Active Directory in order to have a managed profile. To configure Persona Management, you specify a central repository, without changing the user's properties in Active Directory. With this central repository, you can manage a user's profile in one environment without affecting the physical machines that users might also log on to.
With View Persona Management, if you provision desktops with VMware ThinApp applications, the ThinApp sandbox data can also be stored in the user profile. This data can roam with the user but does not significantly affect logon times. This strategy provides better protection against data loss or corruption.
You can configure View personas at several levels: a single View desktop, a desktop pool, an OU, or all View desktops in your deployment. You can also use a standalone version of View Persona Management on physical computers and virtual machines that are not managed by View.
By setting group policies (GPOs), you have granular control of the files and folders to include in a persona:
- Specify whether to include the local settings folder. For Windows 7 or Windows Vista, this policy affects the AppData\Local folder. For Windows XP, this policy affects the Local Settings folder.
- Specify which files and folders to load at login time. For example: Application Data\Microsoft\Certificates. Within a folder, you can also specify files to exclude.
- Specify which files and folders to download in the background after a user logs in to the desktop. Within a folder, you can also specify files to exclude.
- Specify which files and folders within a user's persona to manage with Windows roaming profiles functionality instead of View Persona Management. Within a folder, you can also specify files to exclude.
As with Windows roaming profiles, you can configure folder redirection. You can redirect the same folders that support redirection with Windows Roaming User Profiles.
Accessing USB Devices Connected to the End Point
Administrators can configure the ability to use USB devices, such as thumb flash drives, VoIP (voice-over-IP) devices, and printers, from a View desktop. This feature is called USB redirection. (It was not used in this study.)
When you use this feature, most USB devices that are attached to the local client system become available from
a menu in View Client. You use the menu to connect and disconnect the devices.
You can specify which types of USB devices end users are allowed to connect to. For composite devices that contain multiple types of devices, such as a video input device and a storage device, you can split the device so that one device (for example, the video input device) is allowed but the other device (for example, the storage device) is not.
USB devices that do not appear in the menu, but are available in a View desktop, include smart card readers and human interface devices such as keyboards and pointing devices. The View desktop and the local computer use these devices at the same time.
Printing from a View Desktop
The virtual printing feature allows end users with View Client on Windows systems to use local or network printers from a View desktop without requiring that additional print drivers be installed in the View desktop.
The location-based printing feature allows you to map View desktops to the printer that is closest to the endpoint client device.
With virtual printing, after a printer is added on a local Windows computer, View adds that printer to the list of available printers on the View desktop. No further configuration is required. For each printer available through this feature, you can set preferences for data compression, print quality, double-sided printing, color, and so on. Users who have administrator privileges can still install printer drivers on the View desktop without creating a conflict with the virtual printing component. To send print jobs to a USB printer, you can either use the USB redirection feature or use the virtual printing feature.
The location-based printing feature is available for both Windows and non-Windows client systems. Location based printing allows IT organizations to map View desktops to the printer that is closest to the endpoint client device. Using this feature does require that the correct printer drivers be installed in the View desktop.
We did not use virtual printing in this study. Our workload generator, Login VSI, installs a pdf printer into the master image which is utilized for printing during the test.
Other Features to Consider
Horizon View 5.3 supports these additional features that were not deployed in this study:
Designing a VMware Horizon View 5.3 Deployment
There are several elements that go into the design of a successful Horizon View 5.3 environment. This section covers those topics at a high level. Readers should consult the VMware View Architecture Planning guide for Horizon View 5.3 at the following URL for more details:
https://www.vmware.com/support/view53/doc/horizon-view-53-release-notes.html
Determine Desktop Pools Required
Based on the analysis performed on user groups and the applications identified that will be supported by the Hosted Virtual Desktop (HVD) environment, a strategy for laying out your desktop pool structure should be create.
For this study, we will test a single user group (knowledge workers) and have identified the application workload this group will run, which is based on the Login VSI 3.6 medium workload (with flash.) We will also use virtual machines as our desktop source.
If you use a vSphere virtual machine as a desktop source, you can automate the process of making as many identical virtual desktops as you need. You can set a minimum and maximum number of virtual desktops to be generated for the pool. Setting these parameters ensures that you always have enough View desktops available for immediate use but not so many that you overuse available resources.
Using pools to manage desktops allows you to apply settings or deploy applications to all virtual desktops in a pool. The following examples show some of the settings available:
- Specify which remote display protocol to use as the default for the View desktop and whether to let end users override the default.
- Configure the display quality and bandwidth throttling of Adobe Flash animations.
- If using a virtual machine, specify whether to power off the virtual machine when it is not in use and whether to delete it altogether.
- If using vSphere 4.1 or later, specify whether to use a Microsoft Sysprep customization specification or QuickPrep from VMware. Sysprep generates a unique SID and GUID for each virtual machine in the pool.
- Specify whether the View desktop can or must be downloaded and run on a local client system.
In addition, using desktop pools provides many conveniences.
- Dedicated-assignment pools: Each user is assigned a particular View desktop and returns to the same virtual desktop at each login. Users can personalize their desktops, install applications, and store data.
- Floating-assignment pools: The virtual desktop is optionally deleted and re-created after each use, offering a highly controlled environment. A floating-assignment desktop is like a computer lab or kiosk environment where each desktop is loaded with the necessary applications and all desktops have access to necessary data.
Using floating-assignment pools also allows you to create a pool of desktops that can be used by shifts of users. For example, a pool of 100 desktops could be used by 300 users if they worked in shifts of 100 users at a time.
For this study, we used Automated Pools with Floating Assignments in conjunction with View Composer linked clones.
Managing Storage Requirements
VMware vSphere lets you virtualize disk volumes and file systems so that you can manage and configure storage without having to consider where the data is physically stored.
Fibre Channel SAN arrays, iSCSI SAN arrays, and NAS arrays are widely used storage technologies supported by VMware vSphere to meet different data center storage needs. The storage arrays are connected to and shared between groups of servers through storage area networks. This arrangement allows aggregation of the storage resources and provides more flexibility in provisioning them to virtual machines.
With View 4.5 and later and vSphere 4.1 and later, you can now also use the following features:
- vStorage thin provisioning, which lets you start out with as little disk space as necessary and grow the disk to add space later
- Tiered storage, which allows you to distribute virtual disks in the View environment across high performance storage and lower-cost storage tiers, to maximize performance and cost savings
- Local storage on the ESX/ESXi host for the virtual machine swap files in the guest operating system.
With Horizon View 5.3 and later and vSphere 5.0 and later, you can now also use the following features:
- With the View storage accelerator feature, you can configure ESXi hosts to cache virtual machine disk data.
Using this content-based read cache (CBRC) can reduce IOPS and improve performance during boot storms, when many desktops start up and run anti-virus scans at the same time. Instead of reading the entire OS from the storage system over and over, a host can read common data blocks from cache.- You can deploy a desktop pool on a cluster that contains up to 32 ESXi hosts, but you must store the replica disks on NFS datastores.
Although replica disks must be stored on NFS datastores, OS disks and persistent disks can be stored on NFS or VMFS datastores.
5.3.2.1 View Composer
Because View Composer creates desktop images that share virtual disks with a base image, you can reduce the required storage capacity by 50 to 90 percent.
View Composer uses a base image, or parent virtual machine, and creates a pool of up to 1,000 linked-clone virtual machines. Each linked clone acts like an independent desktop, with a unique host name and IP address, yet the linked clone requires significantly less storage.
When creating a linked-clone desktop pool, a full clone is first made from the parent virtual machine. The full clone, or replica, and the clones linked to it can be placed in a variety of locations. The options are:
- Replica and linked clones on same datastore
- Replica and lined clones on different datastores- As an example, you could place the replicas on low capacity read optimized drives with IOPS
- And place the linked clones on traditional spinning media
- Disposable Disks for Paging and Temp Files- Guest OS page files and temp files are placed here. When the HVD is powered off, this disk is deleted
- Persistent disks for dedicated desktops - End user’s application data and profiles are stored here. The data survives refresh, recompose and rebalance operations.
- Local datastores for floating or stateless desktops - Host local drives store linked clone files, presenting some advantages and several disadvantages. Use this option with care after considering your requirements.
Note For this study, we utilized the replicas and linked clones on different datastores technique.
Hosted Virtual Desktop Infrastructure Design
To implement our automated pool floating desktop delivery model for this study, we followed the VMware View Reference Architecture for virtual desktop delivery.
Figure 7 View Desktop Infrastructure
Learn more about VMware Horizon View planning and design at the following location:
https://www.vmware.com/support/view53/doc/horizon-view-53-release-notes.html
Desktop Delivery Base Image Creation
Create base Windows SP1 Virtual Machine
1.
Select ESXi host in Infrastructure cluster and create a virtual machine to use as Golden Image with windows 7 OS. We used windows 7 32 bit OS for our testing.
For the virtual machine following parameters were used:
–
2.
Figure 8 Windows 7 Golden Image
4.
b.
c.
6.
7.
Optimization of base windows 7 SP1 virtual machine
To optimize windows 7 SP1 32 bit virtual machine, see www.vmware.com/files/pdf/VMware-View-OptimizationGuideWindows7-EN.pdf
Install View 5.3 Virtual Desktop Agent software
1.
2.
4.
5.
Figure 13 Ready to Install the Program
Install additional software
1.
3.
Create a Snapshot for Virtual Machine
1.
2.
3.
Figure 14 Virtual Machine Snapshot
Create customization specification or virtual desktops
1.
2.
3.
5.
Figure 16 Registration Information
6.
7.
8.
Figure 19 Administrator Password Configuration
9.
Figure 20 Time Zone Configuration
10.
Figure 21 Network Configuration
11.
Figure 22 Workgroup or Domain Configuration
12.
Figure 23 Operating System Options
13.
To edit or modify customization specification:
1.
2.
3.
Solution Validation
This section details the configuration and tuning that was performed on the individual components to produce a complete, validated solution.
Configuration Topology for Scalable VMware View 5.3 Virtual Desktop Infrastructure on Cisco UCS and EMC Storage
Figure 26 illustrates the architectural topology for the purpose of this study. The architecture is divided
- Cisco UCS Compute Platform
- The Virtual Desktop Infrastructure that runs on Cisco UCS blade hypervisor hosts
- Network Access layer and LAN
- Storage Access Network (SAN) and EMC VNX Storage array
Figure 26 Cisco Unified Computing System VDI configuration
Figure 27 details the physical configuration of the 2000 seat View 5.3 environment.
Figure 27 Detailed Architecture of the Configuration
Cisco Unified Computing System Configuration
This section talks about the UCS configuration that was done as part of the infrastructure build out. The racking, power and installation of the chassis are described in the install guide and it is beyond the scope of this document. For more details, refer the following documents.
- Cisco UCS 5108 Server Chassis Installation Guide - http://www.cisco.com/en/US/docs/unified_computing/ucs/hw/chassis/install/ucs5108_install.html
- Cisco UCS CLI Configuration guide - http://www.cisco.com/en/US/docs/unified_computing/ucs/sw/cli/config/guide/2.2/b_UCSM_CLI_Configuration_Guide_2_2.html
- Cisco UCS-M GUI Configuration guide - http://www.cisco.com/en/US/docs/unified_computing/ucs/sw/gui/config/guide/2.2/b_UCSM_GUI_Configuration_Guide_2_2.html
Base UCS System Configuration
To configure the Cisco Unified Computing System,
1.
2.
3.
4.
5.
GUI Methodology
1.
2.
3.
4.
Note If the firmware is not current, follow the installation and upgrade guide to upgrade the Cisco UCS Manager firmware. Use the UCS Policy in Service Profiles later in this document to update all UCS components in the solution.
Note The BIOS and Board Controller version numbers do not track the IO Module, Adapter, and CIMC controller version numbers in the packages.
5.
Figure 29 Configure and Enable Server Ports
6.
Figure 30 Configure and Enable Uplink Ethernet Ports
7.
Figure 31 Configure and Enable FC Uplink Ports
Note Use the Configure Unified Ports, Configure Expansion Module Ports to configure FC uplinks. In this example, we configured six FC ports, two of which are in use.
8.
Figure 32 Configure Port Channels and Uplink Interfaces
9.
10.
11.
Figure 35 Create Pools in LAN Tab
Note From the LAN tab in the navigator, under the Pools node, we created a MAC address pool of sufficient size for the environment. In this project, we created a single pool with two address ranges for expandability.
12.
Figure 36 Create Pools in SAN Tab
13.
14.
Figure 38 Create Server UUID Pool
15.
Figure 39 Create Two Server Pools
16.
Figure 40 Create Two Server Pool Policy Qualifications
17.
Figure 41 Create Corresponding Server Pool Policies
18.
Figure 42 Create Virtual Host Bus Adapter Templates
Note Create at least one HBA template for each Fabric Interconnect if block storage will be used. We used the WWPN pool created earlier and the QoS Policy created in the section below.
19.
Note In this project we utilized six VLANs to accommodate our four ethernet system classes, a separate VLAN for infrastructure services, and two VLANs for Nexus 1000V packet and control functions. (N1KV management and VMware Management shared VLAN 164.) We did not use VLAN 166 in the FC variant we deployed in this study. However, if the NFS or iSCSI protocols were used, that VLAN is in place.
20.
Figure 44 Configure vNIC Templates
a.
Figure 45 Create vNIC templates
21.
22.
Figure 46 Settings for the B200 M3 Blades
Figure 47 Advanced Tab - Processor Settings
Figure 48 Advanced Tab - Intel Directed IO Settings
Figure 49 Advanced Tab - RAS Memory Settings
The remaining Advanced tab settings are at platform default or not configured. Similarly, the Boot Options and Server Management tab settings are at defaults.
Note Ensure to click Save Changes at the bottom of the page to preserve this setting. Be sure to add this policy to your blade service profile template.
23.
Figure 50 Create Host Firmware Package Polices
Note We did not use any C-series rack server in this study.
Figure 51 Use of Blade Servers Only
Note Management Firmware Packages are no longer supported. Host Firmware packages replaced this functionality in Cisco UCS Manager 2.1.
24.
Figure 52 Create Service Profile Template
Note In this study, we created two templates, one for each of the UCS Blade Server models used.
Follow through each section, utilizing the policies and objects you created earlier, then click Finish .
Note On the Operational Policies screen, select the appropriate performance BIOS policy you created earlier to insure maximum LV DIMM performance.
Note For automatic deployment of service profiles from your template(s), you must associate a server pool that contains blades with the template.
25.
Figure 53 Identify Service Profile Template
a.
Note We used the default Local Storage configuration in this project. Local drives on the blades were not used.
b.
–
–
c.
d.
e.
Figure 59 Create remaining vNICs
g.
h.
i.
j.
k.
Figure 64 Create Operational Policies
l.
Figure 65 Service Profile Template for the UCS Blade Server B200 M
26.
27.
28.
Figure 66 Create Service Profiles from Template
29.
Figure 67 Enter the naming prefix and the number of Service Profiles
Note Cisco UCS Manager created the requisite number of profiles and because of the Associated Server Pool and Server Pool Qualification policy, the B200 M3 blades in the test environment began automatically associating with the proper Service Profile.
30.
Figure 68 Create B250M2-INFRA-SAN Template
31.
The UCS Blade Servers are ready for hypervisor installation.
QoS and CoS in Cisco Unified Computing System
Cisco Unified Computing System provides different system class of service to implement quality of service including:
- System classes that specify the global configuration for certain types of traffic across the entire system
- QoS policies that assign system classes for individual vNICs
- Flow control policies that determine how uplink Ethernet ports handle pause frames.
Applications like the Cisco Unified Computing System and other time sensitive applications have to adhere to a strict QOS for optimal performance.
System Class Configuration
Systems Class is the global operation where entire system interfaces are with defined QoS rules.
- By default system has Best Effort Class and FCoE Class - Best effort is equivalent in MQC terminology as “match any”
–
–
- Max MTU per Class allowed is 9217.
- Via UCS we can map one CoS value to particular class.
- Apart from FCoE class there can be only one more class can be configured as no-drop property.
- Weight can be configured based on 0 to 10 numbers. Internally system will calculate the bandwidth based on following equation (there will be rounding off the number).
% b/w shared of given Class = Weight of the given priority * 100 / Sum of weights of all priority
Cisco UCS System Class Configuration
Enable QOS on the Cisco Unified Computing System
Here four UCS QoS System Classes have been used to prioritize four types of traffic in the infrastructure:
1.
Figure 70 UCS QoS System Class Configuration
2.
Figure 71 UCS QoS Policy Configuration
3.
Figure 72 Utilize QoS Policy in vNIC Template
This is a unique value proposition for UCS with respect to end-to-end QoS. For example, there is a VLAN for the EMC storage, configure Platinum policy with Jumbo frames and get an end-to-end QoS and performance guarantees from the Blade Servers to the Nexus 1000V virtual distributed switches running in vCenter through the Nexus 5548UP access layer switches.
LAN Configuration
The access layer LAN configuration consists of a pair of Cisco Nexus 5548s (N5Ks,) a family member of our low-latency, line-rate, 10 Gigabit Ethernet and FCoE switches for our VDI deployment.
UCS Connectivity
Four 10 Gigabit Ethernet uplink ports are configured on each of the Cisco UCS 6248 fabric interconnects, and they are connected to the Cisco Nexus 5548 pair in a bow tie manner as shown below in a port channel.
The 6248 Fabric Interconnect is in End host mode, and access to both the Fiber Channel and Ethernet (NAS) data is enabled as per the recommended best practice of the Cisco Unified Computing System. This has been built for scale and have provisioned more than 40 G per Fabric interconnect (Figure 32).
Note The upstream configuration is beyond the scope of this document; There are document [4] available that describes best practices to use the Cisco Nexus 5000 and 7000 Series Switches. There is Layer 3 module newly available with the Nexus 5500 series. However, there is no information regarding the same in this document as it was not used in the studies.
Figure 73 Ethernet Network Configuration with Upstream Cisco Nexus 5500 Series from the Cisco Unified Computing System 6200 Series Fabric Interconnects
EMC VNX5600 LAN Connectivity
The Cisco Nexus 5548UP is used to connect to the EMC VNX 5600 storage system for Fiber Channel and file-based access.
The VNX5600 is equipped with dual-port 8GB FC modules on each controller. These are connected to the pair of Nexus 5548 unified ports to provide block storage access to the environment. (See “SAN Configuration” section.)
The VNX5600 supports two dual-port 10G Data Movers which are connected to the pair of N5Ks downstream. One of the Data Movers is set to Active, with the second providing failover capability. This allows end-to-end 10G access for file-based storage traffic. We have implemented jumbo frames on the ports and have priority flow control on, with Platinum CoS and QoS assigned to the vNICs carrying the storage data access on the Fabric Interconnects. (This configuration was not used in this study, but is shown as a supported option.)
The EMC ethernet connectivity diagram is shown below. There is a total of 40 Gbps bandwidth available for the servers.
Figure 74 EMC VNX Ethernet Connectivity
Figure 75 EMC VNX Fibre Channel Connectivity
For information on configuring ethernet connectivity on a EMC VNX5600 Storage System, refer to the EMC website.
Nexus 1000V Configuration in L3 mode.
1.
2.
3.
The command launches the Nexus 1000V Installation Management Center.
Figure 77 Nexus 1000V Installation Management Center
4.
5.
6.
7.
Figure 80 Select VSM Datastore
8.
a.
b.
c.
Figure 81 Configure L3 mode of installation
9.
10.
11.
12.
Figure 83 Review Configuration
13.
Figure 84 Review Configuration
14.
15.
Enter the following configuration commands, one per line. End with CNTL/Z.
16.
17.
18.
These port-profiles are created by default and do not make any changes.
19.
20.
21.
22.
23.
24.
25.
28.
It will bring up ESXi hosts which are not part of existing configuration.
29.
Figure 86 Select Hosts and Physical Adapters
30.
Figure 87 Select Hosts and Physical Adapters
31.
Figure 88 Network Connectivity
32.
Figure 89 Network Connectivity
33.
Figure 90 Virtual Machine Networking
34.
Note This invokes VMware Update Manger (VUM) to automatically push the VEM installation for the selected ESXi hosts. After successful staging, install and remediation process. The ESXi host will be added to N1KV VSM. From the vCenter task manager, quickly check the process of VEM installation.
In the absence of Update manager: Upload vib file cross_cisco-vem-v162-4.2.1.2.2.1a.0-3.2.1.vib for VEM installation to local or remote datastore which can be obtained by browsing to the management IP address for N1KV VSM Login to ESXi host using ESXi shell or SSH session.
35.
Further to verify putty into N1KV VSM. Run sh module command which will show all the ESXi hosts attached to that VSM.
SAN Configuration
The same pair of Nexus 5548UP switches were used in the configuration to connect between the FC ports on the EMC VNX5600 and the FC ports of the UCS 6248 Fabric Interconnects.
Boot from SAN Benefits
Booting from SAN is another key feature which helps in moving towards stateless computing in which there is no static binding between a physical server and the OS / applications it is tasked to run. The OS is installed on a SAN LUN and boot from SAN policy is applied to the service profile template or the service profile. If the service profile were to be moved to another server, the pwwn of the HBAs and the Boot from SAN (BFS) policy also moves along with it. The new server now takes the same exact character of the old server, providing the true unique stateless nature of the UCS Blade Server.
The key benefits of booting from the network:
- Reduce Server Footprints: Boot from SAN alleviates the necessity for each server to have its own direct-attached disk, eliminating internal disks as a potential point of failure. Thin diskless servers also take up less facility space, require less power, and are generally less expensive because they have fewer hardware components.
- Disaster and Server Failure Recovery. All the boot information and production data stored on a local SAN can be replicated to a SAN at a remote disaster recovery site. If a disaster destroys functionality of the servers at the primary site, the remote site can take over with minimal downtime.
- Recovery from server failures is simplified in a SAN environment. With the help of snapshots, mirrors of a failed server can be recovered quickly by booting from the original copy of its image. As a result, boot from SAN can greatly reduce the time required for server recovery.
- High Availability: A typical data center is highly redundant in nature - redundant paths, redundant disks and redundant storage controllers. When operating system images are stored on disks in the SAN, it supports high availability and eliminates the potential for mechanical failure of a local disk.
- Rapid Redeployment: Businesses that experience temporary high production workloads can take advantage of SAN technologies to clone the boot image and distribute the image to multiple servers for rapid deployment. Such servers may only need to be in production for hours or days and can be readily removed when the production need has been met. Highly efficient deployment of boot images makes temporary server usage a cost effective endeavor.
- Centralized Image Management: When operating system images are stored on networked disks, all upgrades and fixes can be managed at a centralized location. Changes made to disks in a storage array are readily accessible by each server.
- With Boot from SAN, the image resides on a SAN LUN and the server communicates with the SAN through a host bus adapter (HBA). The HBAs BIOS contain the instructions that enable the server to find the boot disk. All FC-capable Converged Network Adapter (CNA) cards supported on Cisco UCS B-series blade servers support Boot from SAN.
- After power on self-test (POST), the server hardware component fetches the boot device that is designated as the boot device in the hardware BOIS settings. Once the hardware detects the boot device, it follows the regular boot process.
Configuring Boot from SAN Overview
There are three distinct phases during the configuration of Boot from SAN. The high level procedures are:
1.
2.
3.
In each of the following sections, each high level phase will be discussed.
SAN Configuration on Nexus 5548UP
The FCoE and NPIV feature has to be turned on in the Nexus 5500 series switch. Make sure you have 8GB SFP+ modules connected to the Nexus 5548UP ports. The port mode is set to AUTO as well as the speed is set to AUTO. Rate mode is “dedicated” and when everything is configured correctly you should see something like the output below on a Nexus 5500 series switch for a given port (for example, Fc1/17).
A Nexus 5500 series switch supports multiple VSAN configurations. A single VSAN was deployed in this study.
Cisco Fabric Manager can also be used to get a overall picture of the SAN configuration and zoning information. As discussed earlier, the SAN zoning is done upfront for all the pwwns of the initiators with the EMC VNX 5600 target pwwns.
The FC connection was used for configuring boot from SAN for all of server blades. In addition, a general purpose 1TB infrastructure LUN for infrastructure virtual machine storage and 16 write-cache LUNs for each VDI host were provisioned.
Single vSAN zoning was set up on the Nexus 5548’s to make LUNs visible to the infrastructure and test servers.
An example SAN zone configuration is shown below on the Fabric A side:
Where 20:00:00:25:b5:c1:00:af /20:00:00:25:b5:c1:00:9f are blade servers pwwn’s of their respective Converged Network Adapters (CNAs) that are part of the Fabric A side.
The EMC FC target ports are 50:06:01:67:08:60:07:FC /50:06:01:6f:08:60:07:FC and belong to one port on the FC modules on SP-A and SP-B.
Similar zoning is done on the second Nexus 5548 in the pair to take care of the Fabric B side as shown below.
Where 20:00:00:25:b5:c1:00:bf /20:00:00:25:b5:c1:00:8f are blade servers pwwn’s of their respective Converged Network Adapters (CNAs) that are part of the Fabric B side.
The EMC FC target ports are 50:06:01:66:08:60:07:FC /50:06:01:6e:08:60:07:FC and belong to the other port on the FC modules on SP-A and SP-B. They were spread across the two controllers for redundancy as shown in the figure below.
Figure 93 VNX5600 FC Target Ports
For detailed Nexus 5500 series switch configuration, refer to Cisco Nexus 5500 Series NX-OS SAN Switching Configuration Guide. (See the Reference Section of this document for a link.)
Configuring Boot from SAN on EMC VNX
The steps required to configure boot from SAN LUNs on EMC VNX are as follow:
1.
Figure 94 Create Storage Pool - General Tab
2.
3.
Figure 96 Create Storage Group
4.
Figure 97 Create Initiator Record
5.
Figure 98 Storage Group Properties
6.
Figure 99 Storage Group Properties
When the UCS Blade Server boots up, its vHBAs will connect to the provisioned EMC Boot LUNs and the hypervisor operating system can be installed.
SAN Configuration on Cisco UCS Manager
To enable Boot from SAN on the Cisco UCS Manager 2.1 (UCS-M) series, do the following:
1.
2.
3.
4.
Figure 101 Create Boot Policy- Add SAN Boot Target
5.
6.
7.
The Boot from SAN policy results in a view as below.
Figure 102 Boot from SAN policy resulting view
8.
EMC VNX5600 Storage Configuration
The following core storage diagram illustrates the layout of the disks that are required to store 2,000 desktop virtual machines. This layout does not include space for user profile data.
Figure 103 Core storage layout
The following core configuration is used in the reference architecture:
- Four SAS disks (0_0_0 to 0_0_3) are used for the VNX OE
- Disks 0_0_6, 0_0_7, and 1_0_2 are hot spares. These disks are marked as hot spare in the storage layout diagram.
- Thirty SAS disks (0_0_10 to 0_0_14, 1_0_5 to 1_0_14, and 0_1_0 to 0_1_14) in the RAID 5 storage pool 0 are used to store virtual desktops. FAST Cache is enabled for the entire pool.
–
–
- Four Flash drives (0_0_4 to 0_0_5 and 1_0_0 to 1_0_1) are used for EMC VNX FAST Cache. There are no user-configurable LUNs on these drives.
- Disks 0_0_8 to 0_0_9 and 1_0_3 to 1_0_4 were not used for testing this solution
In solution validation testing, storage space for user data and infrastructure was allocated on the VNX array as shown in the following figure. This storage is in addition to the core storage shown above. If storage for user data exists elsewhere in the production environment, this storage is not required.
EMC Storage Configuration for VMware View
EMC Storage Configuration for VMware ESXi 5.5 Infrastructure Clusters
The steps required to configure LUNs for the VDI datastores are as follow:
1.
Select Storage > Storage Configuration > Storage Pools . Click Create .
2.
Two LUNs of 2.08TB are derived from the RAID 5 storage pool configured with 10 SAS drives. The LUNs are used to store infrastructure virtual machines such as Active Directory, SQL Server, View Horizon 5.3 components and VMware vCenter server.
Example EMC Boot LUN Configuration
Each ESXi server requires a boot LUN from SAN for the hypervisor OS. A total of 43 LUNs are derived from the 5-disk RAID 5 pool. Each LUN is 5GB in size.
EMC FAST Cache in Practice
EMC FAST Cache uses Flash drives to add an extra layer of cache between the dynamic random access memory (DRAM) cache and rotating disk drives, thereby creating a faster medium for storing frequently accessed data. FAST Cache is an extendable Read/Write cache. It boosts application performance by ensuring that the most active data is served from high-performing Flash drives and can reside on this faster medium for as long as is needed.
FAST Cache tracks data activity at a granularity of 64KB and promotes hot data in to FAST Cache by copying from the hard disk drives (HDDs) to the Flash drives assigned to FAST Cache. Subsequent IO access to that data is handled by the Flash drives and is serviced at Flash drive response times-this ensures very low latency for the data. As data ages and becomes less active, it is flushed from FAST Cache to be replaced by more active data.
Only a small number of Flash drives are needed to enable FAST Cache that provides greater performance instead of a large number of short-stroked HDDs. This results in saving of cost in data center space, power, and cooling requirements that lowers overall TCO for the business.
FAST Cache is particularly suited to applications that randomly access storage with high frequency, such as Oracle and SQL OLTP databases. OLTP databases have inherent locality of reference with varied IO.g
Cisco UCS Manager Configuration for VMware ESXi 5.5
This section addresses creation of the service profiles and VLANs to support the project.
Service Profile Templates
Two types of service profiles were required to support two different blade server types:
To support those different hardware platforms, service profile templates were created, utilizing various policies created earlier.
The service profile templates were then used to quickly deploy service profiles for each blade server in the UCS system. When each blade server booted for the first time, the service profile was deployed automatically, providing the perfect configuration for the VMware ESXi 5.5 installation.
VLAN Configuration
In addition, to control network traffic in the infrastructure and assure priority to high value traffic, virtual LANs (VLANs) were created on the Nexus 5548s, on the Cisco UCS Manager (Fabric Interconnects,) and on the Nexus 1000V Virtual Switch Modules in each vCenter Cluster. The virtual machines in the environment used the VLANs depending on their role in the system.
A total of seven Virtual LANs, VLANs, were utilized for the project. The following table identifies them and describes their use:
To configure VLANs click the LAN/VLANs node in the left pane under LAN tab of Cisco UCS Manager. For more information, see “Base UCS System Configuration” section.
Installing and Configuring ESXi 5.5
Fibre Channel storage are used to boot the hosts from LUNs on the VNX5600 storage system. Prior to installing the operating system, create storage groups and assign specific boot LUNs to individual hosts. (For more information see “Configuring Boot from SAN on EMC VNX” section)
VMware ESXi 5.5 can be installed in boot-from-SAN mode using standard hypervisor deployment techniques including:
1.
2.
Install VMware ESXi 5.5
ESXi was installed using the Cisco UCS Manager KVM console with the Cisco Customized ESXi 5.5 ISO image mounted. The Cisco UCS Manager boot policy deployed to each blade was set to boot from CD then SAN to accommodate hypervisor installs or updates.
The IP address, hostname, and NTP server were configured using Direct Console ESXi Interface accessed from UCSM KVM console ( http://pubs.vmware.com/vsphere-50/topic/com.vmware.vsphere.install.doc_50/GUID-26F3BC88-DAD8-43E7-9EA0-160054954507.html ).
Install and Configure vCenter
To manage hypervisors and virtual machines a dedicated vCenter server instance is installed on Windows 2008R2 SP1 based virtual machine.
To support vCenter instance, create a Microsoft SQL Server 2008 R2 server to host vCenter database. For more information, refer Microsoft documentation on configuring SQL Server and SQL Server clusters. ( http://msdn.microsoft.com/en-us/library/ms189134(v=sql.105).aspx and http://msdn.microsoft.com/en-us/library/ms189134(v=sql.105).aspx)
Install and configure vCenter
1.
2.
3.
4.
5.
6.
Install VMware Licenses
1.
2.
3.
ESXi 5.5 Cluster Configuration
To accommodate maximum recommendations for View 5.3 on ESXi 5.5 we created three ESXi 5.5 clusters described below.
The 16 B200 M3 and 4 B250 M2 ESX hosts were configured into 3 Clusters:
Infra-Cluster Infrastructure Cluster
The Infra-Cluster was used to host all of the virtualized servers within the VDA Infrastructure, including two pairs of Nexus 1000V Virtual Switch Manager (VSM) appliances, one for each virtual desktop cluster.
Use two physical UCS B200 M3 hosts in this cluster.
1.
Virtual Desktop Clusters
VDI-CLUSTER cluster was used to host 2000 desktops:
Cluster was configured with a Nexus 1000V high availability distributed virtual switch providing the required network connectivity.
The Nexus 1000V switches was configured to manage networking for ESXi Cluster hosting virtual desktops, working in concert with the UCS Fabric Interconnects and Nexus 5548UP access layer switches to provide end to end Quality of Service for network communications, insuring the highest quality virtual desktop end user experience.
The Nexus 1000V configuration is described in detail in Section 7.3.3 Nexus 1000V Configuration earlier in this document.
Login VSI Launcher Cluster
Use a separate Launcher-CL cluster to host Login Consultants’ LoginVSI launcher VMs and a LoginVSI console VM.
You may host it on the same UCS Domain with dedicated storage.
The Laucher-CL cluster uses the configured Standard vSwitch configured as follows:
Installing and configuring VMware View 5.3
To build a VMware View 5.3 environment, install the following components:
This section outlines the tasks required to build the View 5.3 environment. For more information, see the VMware View Installation guide for View 5.3.
Prerequisites
The following is a list of per-requisites that are required to install View 5.3 components.
–
–
Note that you can combine operating systems within a site.
- A supported Microsoft SQL or Oracle database for vCenter and View Composer databases
- A supported vSphere hypervisor host operating system
- Physical or virtual hardware meeting the following recommended requirements
–
–
–
–
Create SQL Databases for View 5.3
View Manager Installer needs a separate database for View Composer Server and View Server events.
Create Database for View Composer Server
1.
2.
3.
Create Event Database for View Administrator.
Create a Database for View Administrator Events and user with SA authentication.
Install View Manager and components
Download View Manager software from the link given below.
https://my.vmware.com/web/vmware/details?productId=268&downloadGroup=VIEW-512-PREMIERE
Install View Connection server
1.
2.
Figure 110 Installer for VMware View Connection Server
4.
5.
6.
Figure 113 Installation Options
7.
8.
Figure 115 Firewall Configuration
9.
Figure 116 Install View Administrations
10.
Figure 117 User Experience Improvement Program
11.
Note Need to install Flash player plugin v10.3 or higher to use Web Browser for Login to View Administrator.
12.
Figure 118 VMware Horizon View Administrator
Install View Replica Server
1.
2.
Figure 119 Installation Wizard for VMware View Connection Server
4.
5.
6.
Figure 122 Installation Options
7.
Figure 124 Firewall Configuration
Figure 125 Ready to Install the Program
Install View Composer Server
Install the View Composer server on a separate stand-alone server or on the same server that was used for vCenter server installer.
1.
2.
Figure 126 VMware Horizon View Composer
Figure 127 Installer for VMware View Composer
4.
5.
6.
Figure 130 Database Information
7.
Figure 131 VMware View Composer Port Settings
Figure 132 Ready to install the Program
View Administrator Configuration
To configure the View 5.3 system, follow these steps:
1.
3.
4.
Figure 133 VMware Horizon View Administrator
5.
Figure 134 Add vCenter Servers
6.
–
–
–
–
7.
Figure 136 Invalid Certificate Detected
8.
Figure 139 View Composer Domains
11.
12.
Note In this study a 2048MB cache size was used.
13.
14.
15.
Figure 141 Edit Event Database
16.
Install SSL Certificate for View Connection and Replica Server
1.
2.
3.
4.
Figure 145 Add or Remove Snap-ins
5.
Figure 146 Certificates snap in
6.
7.
Figure 148 Add or Remove Snap-ins
8.
9.
10.
Figure 150 Web Server Properties
11.
12.
13.
Figure 151 Request New Certificate
Figure 152 Certificate Enrollment
15.
17.
18.
Figure 154 Certificate Properties
20.
Figure 155 Certificate Properties - Private Key
21.
22.
23.
Figure 156 Add/remove Snap-in…
24.
25.
Figure 158 Certificates snap in
26.
27.
28.
29.
30.
Install View Client on End Points
1.
https://my.vmware.com/web/vmware/info/slug/desktop_end_user_computing/vmware_horizon_view/5_3?rct=j&q=&esrc=s&source=web&cd=1&sqi=2&ved=0CDEQFjAA&url=http://www.vmware.com/go/downloadview&ei=DQTfUsrJBYe9qAGCkoCYCg&usg=AFQjCNEZ7w5_wmQcbT2GDEdoZ5AbVstNUw
2.
3.
4.
5.
6.
Figure 164 Enhanced Single Sign On
Figure 165 Configure Shortcuts
Figure 166 Ready to Install the Program
Configure Content Based Read Cache (CBRC) on View 5.3 Hosts
CBRC was introduced as a feature of vSphere 5. It is a read cache that is particularly useful during boot storms. It becomes an essential configuration for floating assignment View 5.3 Linked Clones.
The CBRC feature provides a per-host RAM-based solution for View desktops. This considerably reduces the read I/O requests that are issued to the storage layer, and also addresses boot storm snags.
CBRC is configured in vCenter by highlighting the host; access the Configuration Tab, Software, and Advanced Settings.
Each ESXi host used for View Desktops we enabled CBRC and increased the CBRC.DCacheMemReserved to 2048.
Figure 167 Advanced Settings - CBRC
These CBRC settings are used in conjunction with the View 5.3 Administrator, View Configuration, Servers, vCenter Server Properties, Host Cashing tab.
In our test environment, we enabled 2GB of CBRC and correspondingly, 2GB of Host Cache in View Administrator. This combination enables the View Storage Accelerator feature.
Storage Configuration for View 5.3 Hosts
On VNX 5600 30 SAS disks with 300 GB capacity were used to create 16 LUNs, each with a capacity of 375 GB. 2 LUNs with capacities of 50 GB each were created to store replica disks.
Each ESXi host in cluster was assigned 16 LUNs as VMFS5 datastores for linked clones and two 50Gb VMFS5 data store to hold the Replica disk intended for Pool 1 or2 of 1000 vms deployment.
Configure the View Desktop Pools and Options.
Desktop Pools are the containment object in View 5.3 Administrator that hold the configuration and the provisioned linked clones in the View environment.
The maximum recommended number of virtual machines in a VMware ESXi cluster is 1000. Therefore, we created two View 5.3 pools with identical settings to match up with ESXi 5.5 cluster for VDI described earlier in this document.
The following sections describe how we configured our View 5.3 environment.
Create the Desktop Pools
1.
3.
Figure 168 Create Desktop Pool
There are three types of Desktop Pool you can create and description for each type is given on the right side of the screen. For testing purpose, Automated Pool is created.
Two User assignment options are available:.
a.
b.
For our test we used Floating user assignments.
5.
6.
7.
Figure 172 Pool Identification
8.
9.
–
–
Note Use {n} to deploy multiple desktops with same naming pattern. In case of name used VM-{n} deployed desktops will be VM-1, VM-2 ….. VM-10
–
For this study, we provisioned the entire desktops up front.
Figure 174 Provisioning Settings
10.
Figure 175 View Composer Disks
11.
Figure 176 Storage Optimization
12.
For testing, 2 Pools with 1000 desktops in each were created. One Pool was created with Cluster1 as host Resource Pool. 8 VMFS5 datastores for Linked Clones and one Replica disk datastore were selected. The second Pool was configured similarly with the remaining 8 VMFS5 datastores for Linked Clones and the remaining Replica VMFS5 datastore for the Replica disk.
13.
Figure 178 Advanced Storage Options
14.
Figure 179 Guest Customization
15.
16.
17.
18.
19.
After pool setting is completed, it will create a replica from the parent VM and start provisioning of the desktops as per the pool settings.
Test Setup and Configurations
In this project, we tested a single UCS B200 M3 blade in a single chassis and fourteen B200 M3 blades in 2 chassis to illustrate linear scalability.
Cisco UCS Test Configuration for Single Blade Scalability
Figure 183 Cisco UCS B200 M3 Blade Server for Single Server Scalability
Hardware Components
- 1 X Cisco UCS B200-M3 (2 X E5-2680v2 @ 2.80 GHz) blade server with 384GB of memory (16 GB X 24 DIMMS @ 1333 MHz) Windows 7 SP1 Virtual Desktop hosts
- 2 X Cisco UCS B200-M3 (2 X 2650v2 @ 2.60 GHz) blade servers with 128 GB of memory (4 GB X 16 DIMMS @ 1866 MHz) Infrastructure Servers
- 4 X Cisco UCS B250 M2- (2 X 5680 @ 3.333 GHz) blade servers with 192 GB of memory (4 GB X 48 DIMMS @ 1333 MHz) Load Generators
- 2 X M81KR (Palo) Converged Network Adapter/Blade (B250 M2)
- 1X VIC1240 Converged Network Adapter/Blade (B200 M3)
- 2 X Cisco Fabric Interconnect 6248UPs
- 2 X Cisco Nexus 5548UP Access Switches
- 1 X EMC VNX System storage array, two controllers, two Datamovers, 2 x dual port 8GB FC cards, 2 x dual port 10 GbE cards, 4 x 200GB Flash Drives for EMC Fast Cache, 30 x 300GB SAS drives for VMFS datastores, 8 x 600GB SAS Drives for Infrastructure and Boot LUNs and 2 x 300GB SAS drives and 1 200GB Flash Drive for hot spares
Cisco UCS Configuration for Two Chassis – Fourteen Blade Test
Figure 184 Two Chassis Test Configuration-14 x B200 M3 Blade Servers
Hardware Components
- 14 X Cisco UCS B200-M3 (2 X E5-2680 v2 @ 2.80 GHz) blade server with 384GB of memory (16 GB X 24 DIMMS @ 1333 MHz) Windows 7 SP1 Virtual Desktop hosts
- 2 X Cisco UCS B200-M3 (2 X 2650 v2 @ 2.60 GHz) blade servers with 128 GB of memory (8 GB X 16 DIMMS @ 1333 MHz) Infrastructure Servers
- 4 X Cisco UCS B250-M2-2 (2 X 5680 @ 3.333 GHz) blade servers with 192 GB of memory (4 GB X 48 DIMMS @ 1333 MHz) Load Generators
- 2 X M81KR (Palo) Converged Network Adapter/Blade (B250 M2)
- 1X VIC1240 Converged Network Adapter/Blade (B200 M3)
- 2 X Cisco Fabric Interconnect 6248UPs
- 2 X Cisco Nexus 5548UP Access Switches
- 1 X EMC VNX System storage array, two controllers, two Datamovers, 2 x dual port 8GB FC cards, 2 x dual port 10 GbE cards, 4 x 200GB Flash Drives for EMC Fast Cache, 30 x 300GB SAS drives for VMFS datastores, 8 x 600GB SAS Drives for Infrastructure and Boot LUNs and 2 x 600GB SAS drives and 1 200GB Flash Drive for hot spares
Testing Methodology and Success Criteria
All validation testing was conducted on-site within the Cisco Labs with joint support from both Cisco and EMC resources.
The testing results focused on the entire process of the virtual desktop lifecycle by capturing metrics during the desktop boot-up, user logon and virtual desktop acquisition (also referred to as ramp-up,) user workload execution (also referred to as steady state), and user logoff for the Hosted VDI model under test.
Test metrics were gathered from the hypervisor, virtual desktop, storage, and load generation software to assess the overall success of an individual test cycle. Each test cycle was not considered passing unless all of the planned test users completed the ramp-up and steady state phases (described below) and unless all metrics were within the permissible thresholds as noted as success criteria.
Three successfully completed test cycles were conducted for each hardware configuration and results were found to be relatively consistent from one test to the next.
Load Generation
Within each test environment, load generators were utilized to put demand on the system to simulate multiple users accessing the Horizon View 5.3 environment and executing a typical end-user workflow. To generate load within the environment, an auxiliary software application was required to generate the end user connection to the Horizon View environment, to provide unique user credentials, to initiate the workload, and to evaluate the end user experience.
In the Hosted VDI test environment, sessions launchers were used simulate multiple users making a direct connection to Horizon View 5.3 via a VMware Horizon View PCoIP protocol connection.
User Workload Simulation LoginVSI from Login Consultants
One of the most critical factors of validating a Horizon View deployment is identifying a real-world user workload that is easy for customers to replicate and standardized across platforms to allow customers to realistically test the impact of a variety of worker tasks. To accurately represent a real-world user workload, a third-party tool from Login Consultants was used throughout the Hosted VDI testing.
The tool has the benefit of taking measurements of the in-session response time, providing an objective way to measure the expected user experience for individual desktop throughout large scale testing, including login storms.
The Virtual Session Indexer (Login Consultants’ Login VSI 3.6) methodology, designed for benchmarking Server Based Computing (SBC) and Virtual Desktop Infrastructure (VDI) environments is completely platform and protocol independent and hence allows customers to easily replicate the testing results in their environment. In this testing, we utilized the tool to benchmark our VDI environment only.
Login VSI calculates an index based on the amount of simultaneous sessions that can be run on a single machine.
Login VSI simulates a medium workload user (also known as knowledge worker) running generic applications such as: Microsoft Office 2007 or 2010, Internet Explorer 8 including a Flash video applet and Adobe Acrobat Reader (For the purposes of this test, applications were installed locally, not streamed nor hosted on Thin App).
Like real users, the scripted Login VSI session will leave multiple applications open at the same time. The medium workload is the default workload in Login VSI and was used for this testing. This workload emulated a medium knowledge working using Office, IE, printing and PDF viewing.
- Once a session has been started the medium workload will repeat every 12 minutes.
- During each loop the response time is measured every 2 minutes.
- The medium workload opens up to 5 apps simultaneously.
- The type rate is 160ms for each character.
- Approximately 2 minutes of idle time is included to simulate real-world users.
- Outlook 2007/2010, browse 10 messages.
- Internet Explorer, one instance is left open (BBC.co.uk), one instance is browsed to Wired.com, Lonelyplanet.com and heavy
- 480 p Flash application gettheglass.com.
- Word 2007/2010, one instance to measure response time, one instance to review and edit document.
- Bullzip PDF Printer & Acrobat Reader, the word document is printed and reviewed to PDF.
- Excel 2007/2010, a very large randomized sheet is opened.
- PowerPoint 2007/2010, a presentation is reviewed and edited.
- 7-zip: using the command line version the output of the session is zipped.
A graphical representation of the medium workload is shown below.
For more information see, http://www.loginvsi.com .
Testing Procedure
The following protocol was used for each test cycle in this study to insure consistent results.
Pre-Test Setup for Single and Multi-Blade Testing
- All virtual machines were shut down utilizing the vCenter
- All Launchers for the test were shut down. They were then restarted in groups of 10 each minute until the required number of launchers was running with the Login VSI Agent at a “waiting for test to start” state.
- All VMware ESXi 5.5 VDI host blades to be tested were restarted prior to each test cycle.
Test Run Protocol
To simulate severe, real-world environments, Cisco requires the log-on and start-work sequence, known as Ramp Up, to complete in 30 minutes. Additionally, we require all sessions started, whether 170 single server users or 2000 full scale test users to become active within 2 minutes after the session is launched.
For each of the three consecutive runs on single blade (170 User) and 14-blade (2000 User) tests, the same process was followed:
1.
–
–
–
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Note 170 users will be available in 2 minutes and 2000 users available in 16 minutes approximately.
Success Criteria
There were multiple metrics that were captured during each test run, but the success criteria for considering a single test run as pass or fail was based on the key metric, VSI Max. The Login VSI Max evaluates the user response time during increasing user load and assesses the successful start-to-finish execution of all the initiated virtual desktop sessions.
Login VSI Max
VSI Max represents the maximum number of users the environment can handle before serious performance degradation occurs. VSI Max is calculated based on the response times of individual users as indicated during the workload execution. The user response time has a threshold of 4000ms and all users response times are expected to be less than 4000ms in order to assume that the user interaction with the virtual desktop is at a functional level. VSI Max is reached when the response times reaches or exceeds 4000ms for 6 consecutive occurrences. If VSI Max is reached, that indicates the point at which the user experience has significantly degraded. The response time is generally an indicator of the host CPU resources, but this specific method of analyzing the user experience provides an objective method of comparison that can be aligned to host CPU performance.
Note In the prior version of Login VSI, the threshold for response time was 2000ms. The workloads and the analysis have been upgraded in Login VSI 3 to make the testing more aligned to real-world use. In the medium workload in Login VSI 3.0, a CPU intensive 480p flash movie is incorporated in each test loop. In general, the redesigned workload would result in an approximate 20% decrease in the number of users passing the test versus Login VSI 2.0 on the same server and storage hardware.
Calculating VSIMax
Typically the desktop workload is scripted in a 12-14 minute loop when a simulated Login VSI user is logged on. After the loop is finished it will restart automatically. Within each loop the response times of seven specific operations is measured in a regular interval: six times in within each loop. The response times if these seven operations are used to establish VSImax. The seven operations from which the response times are measured are:
1.
This operation will refresh a new document to be used for measuring the response time. This activity is mostly a file-system operation.
2.
This operation will measure the responsiveness of the Operating System and the file system. Microsoft Word is started and loaded into memory; also the new document is automatically loaded into Microsoft Word. When the disk I/O is extensive or even saturated, this will impact the file open dialogue considerably.
3.
This operation is handled for small part by Word and a large part by the operating system. The file open dialogue uses generic subsystems and interface components of the OS. The OS provides the contents of this dialogue.
This operation is handled by the OS (loading and initiating notepad.exe) and by the Notepad.exe itself through execution. This operation seems instant from an end-user’s point of view.
5.
This operation is handled for a large part by the OS subsystems, as the print dialogue is provided by the OS. This dialogue loads the print-subsystem and the drivers of the selected printer. As a result, this dialogue is also dependent on disk performance.
6.
This operation is handled within the application completely; the presentation of the dialogue is almost instant. Serious bottlenecks on application level will impact the speed of this dialogue.
7.
This operation is handled by the command line version of 7-zip. The compression will very briefly spike CPU and disk I/O.
These measured operations with Login VSI do hit considerably different subsystems such as CPU (user and kernel), Memory, Disk, the OS in general, the application itself, print, GDI, and so on. These operations are specifically short by nature. When such operations are consistently long: the system is saturated because of excessive queuing on any kind of resource. As a result, the average response times will then escalate. This effect is clearly visible to end-users. When such operations consistently consume multiple seconds the user will regard the system as slow and unresponsive. With Login VSI 3.0 and later it is now possible to choose between ‘VSImax Classic’ and 'VSImax Dynamic’ results analysis. For these tests, we utilized VSImax Dynamic analysis.
VSIMax Dynamic
VSImax Dynamic is calculated when the response times are consistently above a certain threshold. However, this threshold is now dynamically calculated on the baseline response time of the test. Five individual measurements are weighted to better support this approach:
- Copy new doc from the document pool in the home drive: 100%
- Microsoft Word with a document: 33.3%
- Starting the “File Open” dialogue: 100%
- Starting “Notepad”: 300%
- Starting the “Print” dialogue: 200%
- Starting the “Search and Replace” dialogue: 400%
- Compress the document into a zip file with 7-zip command line 200%
A sample of the VSImax Dynamic response time calculation is displayed below:
Then the average VSImax response time is calculated based on the amount of active Login VSI users logged on to the system. For this the average VSImax response times need to consistently higher than a dynamically calculated threshold. To determine this dynamic threshold, first the average baseline response time is calculated. This is done by averaging the baseline response time of the first 15 Login VSI users on the system.
The formula for the dynamic threshold is: Avg. Baseline Response Time x 125% + 3000. As a result, when the baseline response time is 1800, the VSImax threshold will now be 1800 x 125% + 3000 = 5250ms.
Especially when application virtualization is used, the baseline response time can wildly vary per vendor and streaming strategy. Therefore it is recommend to use VSImax Dynamic when comparisons are made with application virtualization or anti-virus agents. The resulting VSImax Dynamic scores are aligned again with saturation on a CPU, Memory or Disk level, also when the baseline response time are relatively high.
Determining VSIMax
The Login VSI analyzer will automatically identify the “VSImax”. In the example below the VSImax is 99. The analyzer will automatically determine “stuck sessions” and correct the final VSImax score.
Figure 186 Sample Login VSI Analyzer Graphic Output
- Red line: Maximum Response (worst response time of an individual measurement within a single session)
- Orange line: Average Response Time within for each level of active sessions
- Blue line: the VSImax average.
- Green line: Minimum Response (best response time of an individual measurement within a single session)
In our tests, the total number of users in the test run had to login, become active and run at least one test loop and log out automatically without reaching the VSI Max to be considered a success.
Note We discovered a technical issue with the VSIMax dynamic calculation in our testing on Cisco B200 M3 blades where the VSIMax Dynamic was not reached during extreme conditions. Working with Login Consultants, we devised a methodology to validate the testing without reaching VSIMax Dynamic until such time as a new calculation is available.
Our Login VSI “pass” criteria, accepted by Login Consultants for this testing follows:
1.
2.
3.
4.
a.
b.
5.
The purpose of this testing is to provide the data needed to validate VMware View 5.3 automated pool,
floating assignment linked clone virtual desktops using ESXi 5.5 and vCenter 5.5 to virtualize Microsoft
Windows 7 SP1 desktops on Cisco UCS B200 M3 blade servers using a EMC VNX5600 storage system.
The information contained in this section provides data points that a customer may reference in
designing their own implementations. These validation results are an example of what is possible under
the specific environment conditions outlined here, and do not represent the full characterization of View 5.3 with VMware vSphere.
Two test sequences, each containing three consecutive test runs generating the same result, were performed to establish single server performance and multi-server, linear scalability.
One additional series of stress tests on a single blade server was conducted to establish the official Login
VSI Max Score. To reach the Login VSI Max, we ran 210 Medium Workload (with flash) Windows 7
SP1 sessions on a single server. The Login VSI score was achieved on three consecutive runs and is
VDI Test Results
The purpose of this testing is to provide the data needed to validate VMware View 5.3 automated pool, floating assignment linked clone virtual desktops using ESXi 5.5 and vCenter 5.5 to virtualize Microsoft Windows 7 SP1 desktops on Cisco UCS B200 M3 blade servers using a EMC VNX 5600 storage system.
The information contained in this section provides data points that a customer may reference in designing their own implementations. These validation results are an example of what is possible under the specific environment conditions outlined here, and do not represent the full characterization of View 5.3 with VMware vSphere.
Two test sequences, each containing three consecutive test runs generating the same result, were performed to establish single server performance and multi-server, linear scalability.
One additional series of stress tests on a single blade server was conducted to establish the official Login VSI Max Score. To reach the Login VSI Max, we ran 210 Medium Workload (with flash) Windows 7 SP1 sessions on a single server. The Login VSI score was achieved on three consecutive runs and is shown below.
Figure 187 Login VSIMax Reached: 210 Users
Cisco UCS Test Configuration for Single-Server Scalability Test Results
This section details the results from the View 5.3 Hosted VDI single blade server validation testing. The primary success criteria used to validate the overall success of the test cycle is an output chart from Login Consultants’ VSI Analyzer Professional Edition, VSIMax Dynamic for the Medium workload (with Flash.)
We did not reach a VSIMax Dynamic in our testing due to a technical issue with the analyzer formula that calculates VSIMax. See Section 8.3.4.5 Determining VSIMax for a discussion of this issue.
We ran the single server test at approximately 10% lower user density than prescribed by the Login VSI Max to achieve a successful pass of the test with server hardware performance in a realistic range.
Additionally, graphs detailing the CPU, Memory utilization and network throughput during peak session load are also presented. Given adequate storage capability, the CPU utilization determined the maximum VM density per blade.
The charts below present our recommended maximum Login VSI Medium workload loading on a single blade server.
Figure 188 Users View 5.3 Desktop Sessions on VMware ESXi 5.5 below 4000 ms
The following graphs detail CPU, Memory, Disk and Network performance on the Single Cisco UCS B200-M3 Blades
Figure 189 170 User Single B200 M3 CPU Core Utilization - Boot Phase
Figure 190 170 Users Single B200 M3 CPU Total Utilization - Boot Phase
Figure 191 170 User Single B200 M3 Available Memory MBytes - Boot Phase
Figure 192 170 User Single B200 M3 Cisco VIC1240 MLOM Mbps Received/Transmitted - Boot Phase
Figure 193 170 User Single B200 M3 Cisco VIC1240 MLOM Physical Disk Adapter Mbps Read/Write - Boot Phase
Figure 194 170 User Single B200 M3 CPU Core Utilization - Test Phase
Figure 195 170 User Single B200 M3 CPU Processor Total Utilization Time - Test Phase
Figure 196 170 User Single B200 M3 Available Memory MBytes Test Phase
Figure 197 170 User Single B200 M3 Cisco VIC1240 MLOM Mbps Read/Written Test Phase
Figure 198 170 User Single B200 M3 Cisco VIC1240 MLOM Mbps Received/Transmitted Test Phase
The following charts detail infrastructure server performance during the single blade, 170 User test:
Figure 199 170 User View Connection Server 5.3 CPU Processor Time - Boot Phase
Figure 200 170 User View Connection Server 5.3 CPU Processor User Time- Boot Phase
Figure 201 170 User View Connection Server 5.3 Available Memory Bytes - Boot Phase
Figure 202 170 User View Connection Server 5.3 Bytes Received /Second - Boot Phase
Figure 203 170 User View Connection Server 5.3 Bytes Sent/Second - Boot Phase
Figure 204 170 User View 5.3 - View Connection Server CPU Processor User- Test Phase
Figure 205 170 User View 5.3 - View Connection Total User Time- TEST Phase
Figure 206 170 User View 5.3 - View Connection Server Available Memory Bytes -Test Phase
Figure 207 170 User View 5.3 View Connection Server Bytes Received-Test Phase
Figure 208 170 User View 5.3 View Connection Server Bytes Sent - Test Phase
Cisco UCS Test Configuration for 2000 Desktop Two-Cluster Scalability Test Results
This section details the results from the View 5.3 Hosted VDI fourteen blade server, one-cluster, 2000 user validation testing. It demonstrates linear scalability for the system. The primary success criteria used to validate the overall success of the test cycle is an output chart from Login Consultants’ VSI Analyzer Professional Edition, VSIMax Dynamic for the Medium workload (with Flash.)
We did not reach a VSIMax Dynamic in our testing due to a technical issue with the analyzer formula that calculates VSIMax. See Section 8.3.4.5 Determining VSIMax for a discussion of this issue.
We ran the multi-server test at an average user density slightly below 143 users per blade across the system. One ESX Cluster, containing 14 B200 M3s ran the entire workload. In fact the fourteen blade test harness provides N+1 server fault tolerance on a system basis to achieve a successful pass of the test with server hardware performance in a realistic range.
Additionally, graphs detailing the CPU, Memory utilization and network throughput during peak session load are also presented for a representative blade running 143 user sessions below. We have provided performance charts for all 14 blades in Appendix B to illustrate this point.
Given adequate storage capability, the blade CPU utilization determined the maximum recommended VM density per blade for the 2000 user environment.
For the large scale test, we are including the EMC VNX 5600 performance metrics as well.
Figure 209 2000 Desktop Sessions on VMware ESXi 5.5 below 4000 ms
The following graphs detail CPU, Memory, Disk and Network performance on a representative Cisco UCS B200 M3 Blade during the fourteen blade, 2000 User test. (Representative results for all fourteen blades in vCenter cluster can be found in Appendix B.)
Figure 210 2000 User Single B200 M3 Core CPU Utilization -Boot Phase
Figure 211 2000 User Single B200 M3 CPU Utilization -Boot Phase
Figure 212 2000 User Single B200 M3 Cisco VIC1240 MLOM Physical Disk Adapter Mbps Read/Write - Boot Phase
Figure 213 2000 User Single B200 M3 Cisco VIC1240 MLOM VIC Mbps Received/Transmitted -Boot Phase
Figure 214 2000 User Single B200 M3 Cisco NonKernel MBytes Available -Boot Phase
Figure 215 2000 User Single B200 M3 CPU Utilization Test Phase
Figure 216 2000 User Single B200 M3 Cisco VIC1240 MLOM Network Adapter Mbps Receive/Transmit -Test Phase
Figure 217 2000 User Single B200 M3 Cisco VIC1240 MLOM Physical Disk Adapter Mbps Read/Write -Test Phase
Figure 218 2000 User Single B200 M3 Cisco VIC1240 MLOM VIC Mbps Received/Transmitted - Test Phase
Figure 219 2000 User Single B200 M3 Cisco NonKernel MBytes Available -Test Phase
The following charts detail the VNX5600 performance during the fourteen blade, 2000 User test:
Figure 220 2000 Users EMC VNX5600 SP Utilization -Boot Phase
Figure 221 2000 Users EMC VNX5600 SP Total Throughput Boot Phase
Figure 222 2000 Users EMC VNX5600 SP Total Bandwidth Boot Phase
Figure 223 2000 Users EMC VNX5600 SP Utilization TEST Phase
Figure 224 2000 Users EMC VNX5600 SP Total Throughput Test Phase
Figure 225 2000 Users EMC VNX5600 SP Total Bandwidth Boot Phase
The following charts detail infrastructure server performance during the fourteen blade, 2000 User test:
Figure 226 2000 User View Connection Server 5.3 Processor Total - Test Phase
Figure 227 2000 User View Connection Server 5.3 Processor Total User Time -Test Phase
Figure 228 2000 User View Connection Server 5.3 Bytes Received/Second - Test Phase
Figure 229 2000 User View Connection Server 5.3 Bytes Sent/Second - Test Phase
Figure 230 2000 User View Connection Server 5.3 Available Memory Bytes -Test Phase
Scalability Considerations and Guidelines
There are many factors to consider when you begin to scale beyond 2000 User, two chassis, 14 VDI host server configuration, which this reference architecture has successfully tested. In this section we give guidance to scale beyond the 2000 user system.
Cisco UCS System Configuration Considerations
As our results indicate, we have proven linear scalability in the Cisco UCS Reference Architecture as tested.
- Cisco UCS 2.2(1.b) management software supports up to 20 chassis within a single Cisco UCS domain on our second generation Cisco UCS Fabric Interconnect 6248 and 6296 models. Our single UCS domain can grow to 160 half-width blades.
- With Cisco UCS 2.2(1.b) management software, released late in December 2013, each UCS 2.2(1.b) Management domain is extensibly manageable by UCS Central, our new manager of managers, vastly increasing the reach of the UCS system.
- As scale grows, the value of the combined UCS fabric, Nexus physical switches and Nexus virtual switches increases dramatically to define the Quality of Services required to deliver excellent end user experience 100% of the time.
- To accommodate the Cisco Nexus 5500 upstream connectivity in the way we describe in the LAN and SAN Configuration section, we need four Ethernet uplinks and two Fibre Channel uplinks to be configured on the Cisco UCS Fabric interconnect. And based on the number of uplinks from each chassis, we can calculate number of desktops can be hosted in a single UCS domain. Assuming eight links per chassis, four to each 6248, scaling beyond 10 chassis would require a pair of Cisco UCS 6296 fabric interconnects. A 20,000 virtual desktop building block, with its support infrastructure services can be built out of the RA described in this study with eight links per chassis and 20 Cisco UCS chassis comprised of seven B200 M3 VDI blade server and one B200 M3 Infrastructure blades servers in each chassis.
Of course, the backend storage has to be scaled accordingly, based on the IOPS considerations as described in the EMCs scaling section. Please refer the EMC section that follows this one for scalability guidelines.
VMware View 5.3 Considerations
VMware View Composer can create and provision up to 1000 desktops per pool when deployed on vSphere 4.1 or later. View Composer can also perform a recompose operation on up to 1,000 desktops at a time. Desktop pool size is limited by the following factors:
- Each desktop pool can contain only one ESX/ESXi cluster.
- With View 5.3 and later and vSphere 5.0 and later, an ESXi cluster can contain more than 8 ESXi hosts (up to 32), but you must store the linked-clone replica disks on NFS datastores.
- Each CPU core has compute capacity for 8 to 10 virtual desktops.
A single VMware View Connection server can host up to 2000 simultaneous connections over any supported connection type. Seven View Connection Servers (5 active plus 2 spares) can host up to 10000 direct, RDP or PCoIP connections simultaneously. The sever View Connection Server cluster configuration should not be clustered across WAN links.
VMware View deployments can use VMware HA clusters to guard against physical server failures. With View 5.3 and later and vSphere 5 and later, if you use View Composer and store replica disks on NFS datastores, the cluster can contain up to 32 servers, or nodes.
With vCenter 4.1 and 5.0, each vCenter Server can support up to 10,000 virtual machines.
For more information on VMware View 5.3 configuration and guidelines, see Chapter 11 References.
EMC VNX Storage Guidelines for Horizon View 5.3 Provisioned Virtual Machines
Sizing VNX storage system to meet virtual desktop IOPS requirement is a complicated process. When an I/O reaches the VNX storage, it is served by several components such as Data Mover (NFS), backend dynamic random access memory (DRAM) cache, FAST Cache, and disks. To reduce the complexity, EMC recommends using a building block approach to scale to thousands of virtual desktops.
For more information on storage sizing guidelines to implement virtual desktop infrastructure in VNX unified storage systems, refer to the EMC white paper “Sizing EMC VNX Series for VDI workload – An Architectural Guideline”.
VMware ESXi 5.5 Guidelines for Virtual Desktop Infrastructure
In our test environment two adjustments were performed to support our scale:
- The amount of memory configured for the Tomcat Maximum memory pool was increased to 3072.
- The cost threshold for parallelism was increased to 15.
For further explanations on a basis for these adjustments and details on how to perform them, refer to the VMware documentation sited in the Chapter 11 References section of this document.
References
This section provides links to additional information for each partner’s solution component of this document.
Cisco Reference Documents
- Third-Generation Fabric Computing: The Power of Unification webcast replay: http://tools.cisco.com/gems/cust/customerSite.do?METHOD=W&LANGUAGE_ID=E&PRIORITY_CODE=215011_15&SEMINAR_CODE=S15897&CAMPAIGN=UCS+Momentum&COUNTRY_SITE=us&POSITION=banner&REFERRING_SITE=go+unified+computing&CREATIVE=carousel+banner+event+replay
- Cisco Unified Computing System Manager Home Page: http://www.cisco.com/en/US/products/ps10281/index.html
- Cisco UCS B200 M3 Blade Server Resources: http://www.cisco.com/en/US/partner/products/ps12288/index.html
- Cisco UCS 6200 Series Fabric Interconnects: http://www.cisco.com/en/US/partner/products/ps11544/index.html
- Cisco Nexus 1000V Series Switches Resources: http://www.cisco.com/en/US/partner/products/ps9902/index.html
- Cisco Nexus 5500 Series Switches Resources: http://www.cisco.com/en/US/products/ps9670/index.html
- Download Driver Software for UCS B200 M3 Blade Server: http://software.cisco.com/download/release.html?mdfid=283853163&flowid=25821&softwareid=283853158&release=2.2.1
- Download Cisco UCS Manager and Blade Software Version 2.2(1b) http://software.cisco.com/download/release.html?mdfid=283612660&flowid=22121&softwareid=283655658&release=2.0%285a%29&relind=AVAILABLE&rellifecycle=&reltype=latest
- Download Cisco UCS Central Software Version 1.1(1a): http://software.cisco.com/download/cart.html?imageGuId=8CAAAD77B3A1DB35B157BE84ED109A4703849F53&i=rs
View 5.3 Documents
- Performance and Best Practices - http://www.vmware.com/files/pdf/view/vmware-horizon-view-best-practices-performance-study.pdf
- View 5.3 Architecture and Planning https://www.vmware.com/support/view53/doc/horizon-view-53-release-notes.html
- https://www.vmware.com/support/pubs/view_pubs.html
- View 5 with PCoIP Network Optimization Guide – http://www.vmware.com/files/pdf/view/VMware-View-5-PCoIP-Network-Optimization-Guide.pdf Virtual Desktop
- Windows 7 Optimization Guide: http://www.vmware.com/files/pdf/VMware-View-OptimizationGuideWindows7-EN.pdf
VMware Reference Documents
- Accessing a vCenter Server using Web access or vSphere Client fails with an SSL certificate error: http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1021514
- VMware vSphere ESXi and vCenter Server 5 Documentation: http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1021514
- VMware vCenter Management Webservices features do not function properly: - http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1039180
- VMware® vCenter Server™ 5.5 Database Performance Improvements and Best Practices for Large-Scale Environments: - Performance Best Practices for VMware vSphere™ 5.5: http://www.vmware.com/pdf/Perf_Best_Practices_vSphere5.5.pdf
Feedback