Table Of Contents
Cisco Application Networking for Microsoft SharePoint Solutions Deployment Guide
February 18, 2009
To address challenges associated with today's mission critical enterprise application deployments, Cisco, in collaboration with Microsoft, offers Cisco Application Networking Solution for Microsoft SharePoint Solutions, an enterprise network architecture for deploying Microsoft Office SharePoint Server (MOSS) 2007 with Cisco networking equipment (Joint Solution) with best practices and implementation guidance that optimizes application availability, performance, and security and lowers application ownership costs.
Featuring the Cisco Application Control Engine (ACE) and Wide Area Application Services (WAAS) product families, collectively known as Cisco Application Networking Services (ANS), that provide data center, branch, and remote end user application optimization services, the Joint Solution addresses the following challenges:
•Recovery time and point objectives for business continuity
•Application response time over limited Wide Area Network (WAN) connections
•Application, server, network, and service oriented architecture (SOA) security
•Reduced capital and operational costs for applications, servers, and networking
Cisco and Microsoft cooperated in all phases of the Cisco Application Networking for Microsoft SharePoint Solutions testing, including lab set-up at Cisco and Microsoft offices, solution functional and performance testing, and this deployment guide. Cisco and Microsoft jointly validate that the lab setup and Joint Solution testing represents best efforts in creating a realistic customer deployment and accurate documentation of such deployment.
The purpose of this document is to describe the Joint Solution enterprise network architecture along with deployment best practices and guidance.
The following prerequisites are required to deploy the Joint Solution:
•Working knowledge of the MOSS application
•Experience with basic networking and troubleshooting
•Experience installing the Cisco products covered by this network design, including the Cisco ACE and WAAS product families
•Working knowledge of Cisco's Internetworking Operating System (IOS)
The following table provides a brief description of each section.
The Joint Solution offers optimized MOSS 2007 application availability, performance, security, and costs by providing application optimization services as follows:
Cisco ACE product family application optimization services for MOSS 2007 high availability:
–Server and application health monitoring—Continuously and intelligently monitors application and database availability.
–Server load balancing—Efficiently routes end user and Web services requests to the best available server.
–Network platform health monitoring—Ensures continuity of business operations through mirroring end user transaction states across pairs of network devices.
Cisco ACE and WAAS product family application optimization services for MOSS 2007 high performance:
–WAN optimization—Provides intelligent caching, compression, and protocol optimization that yields up to 34 times faster downloads, 50 percent faster site navigation, and 88 percent less bandwidth (see Results and Conclusions).
–Server offloading—Specialized hardware that offers greater processing efficiency for application optimization services such as server load balancing, Secure Socket Layer termination, and traffic compression, which frees up to 71 percent of application server processing and memory to focus on business logic computations.
–Server load balancing—Substitutes for Microsoft Network Load Balancing (NLB) load balancing.
–Secure Socket Layer (SSL) termination—Terminates 15,000 connections per second.
–Transmission Control Protocol (TCP) connection management—Reduces the number of TCP connections to server.
–Server health monitoring—Substitutes for SharePoint native server health monitoring.
–Traffic compression—Scalable LZ compression functionality.
–Object caching—Reduce requests to server.
Cisco ACE product family application optimization services for MOSS application and data security:
–SSL termination—Efficiently encrypts and decrypts SSL enabled traffic, which facilitates the use of intrusion detection and prevention solutions before traffic reaches the servers.
–End user access control—Provides Access Control Lists (ACLs) to protect client-to-server traffic from worms and intruders that attack vulnerable open server ports not used by the application.
•Virtualization of application optimization services
Virtualization of application optimization services supplies such services for multiple SharePoint instances as well as other enterprise applications (see Figure 1). Specifically, a single physical Cisco ACE can be virtualized into multiple logical Cisco ACEs in which application traffic can traverse between virtualized Cisco ACEs. This virtualization of load balancing is an exclusive Cisco feature.
Figure 1 Virtualization of Application Optimization Services
The application optimization services of the Joint Solution reside in both the data center and the branch to offer end-to-end value, from branch and remote users, all the way through to the database and information storage.
•Data center application optimization services
Cisco ACE and WAAS reside in the data center and are arranged to provide virtualized application optimization services for multiple MOSS 2007 instances as well as other enterprise applications.
Because of their unique location, these solutions can take intelligent action on end-user traffic before it is routed to the MOSS 2007 application servers, including server load balancing, server health monitoring, SSL decryption, TCP connection consolidation, and end user access control.
While some of these functions could be provided natively by Microsoft or third party server-based solutions, Cisco networking provides these services cost-effectively, freeing up server processing and memory needs to focus on business logic computation.
•Wide area application optimization services
Cisco WAAS also resides in the branch office and is arranged to provide virtualized application optimization services for all application users in that location. Together with the data center WAAS deployment, the two offer a WAN optimization service through the use of intelligent caching, compression, and protocol optimization.
When the MOSS 2007 application servers respond to end-user requests, Cisco WAAS compresses the response and then most efficiently passes it across the WAN with minimal bandwidth usage and maximum speed. Commonly used information is cached both at the WAAS solution in the branch as well as in the Cisco ACE solution in the data center, which significantly reduces the burden on the servers and the WAN.
Note The Cisco Wide Area Application Services (WAAS) software runs on the Cisco Wide-Area Application Engine (WAE).
Figure 2 Process Flow
SharePoint Application Overview
MOSS 2007 is a portal-based collaboration and document management platform. It can be used to host web sites, termed SharePoint Portals, which can be used to access shared workspaces and documents, as well as specialized applications, such as wikis and blogs, from within a browser.
MOSS 2007 functionality is exposed as Web parts, which are components that implement certain functionality, such as a task list or discussion pane. These Web parts are then composed into Web pages, which are then hosted in the SharePoint Portal. SharePoint sites are actually ASP.NET applications, which are served using Microsoft Internet Information Server (IIS) and use a Microsoft SQL Server database as data storage backend.
The SharePoint family of products is composed of three different applications. Windows SharePoint Services (WSS) is a free add-on to Windows Server. WSS offers the basic portal infrastructure and collaborative editing of documents, as well as document organization and version control capabilities. It also includes end user functionality such as workflows, to-do lists, alerts, and discussion boards, which are exposed as web parts to be embedded into SharePoint Portal pages. WSS was previously known as SharePoint Team Services.
Microsoft Office SharePoint Server (MOSS) is a paid component of Microsoft Office suite. MOSS integrates with WSS and adds more functionality to it, including better document management, indexed search functionality, navigation features, RSS support, wikis, and blogs, as well as features from Microsoft Content Management Server. It also includes features for business data analysis as well as integration with Microsoft Office applications, such as project management capabilities or exposing Microsoft Office InfoPath forms via a browser. It can also host specific libraries, such as PowerPoint Template Libraries provided the server components of the specific application are installed. MOSS was previously known as SharePoint Server and SharePoint Portal Server.
MOSS 2007 is composed of a three-tier architecture as described below. The first tier is a Web browser for a client. The middle tier consists of a Web and application server running the WSS application with MOSS plugging-in functionality where required, generally a search service which crawls the data store creating an index, and a number of other services. The third tier is the database server.
The middle tier can be scaled out by load balancing more web and application servers in the middle tier and building larger clusters of SQL Server in the third tier.
Application and Application Networking Architecture
Figure 3 Application and Application Networking Architecture
The Joint Solution uses WAAS to enhance performance and Cisco ACE to reduce the load on resources in the server farm. The WAAS and Cisco ACE each provide a unique benefit to the solution, however there are additional benefits when they are used together as the two solutions are complimentary. The Cisco ACE provides load balancing to the server farm. If the application uses SSL, then the Cisco ACE can provide SSL termination offload, thereby increasing efficiency by removing the load on the servers' resources and allowing the servers to process more transactions. Increased server efficiency also results if the Cisco ACE is used to provide TCP reuse.
The Joint Solution architecture is based on the Enterprise Branch Wide Area Application Services Design Guide (Enterprise Branch Design) and the Data Center Infrastructure Design Guide 2.5, both found at http://www.cisco.com/go/designzone.
In the Joint Solution architecture, the WAAS Solution is installed within the Cisco Wide Area Application Engine (WAE) Appliances.
The Enterprise Branch Design shows the Cisco WAE appliance connected to the local branch router, typically a Cisco Integrated Services Router (ISR). The design provides scalability and availability as compared to installing a Cisco WAAS Network Module within a Cisco ISR as the Cisco ISR must share its resources.
HP Mercury LoadRunner, running on a personal computer in the branch, simulates users that would perform certain tasks in the application.
The traffic is redirected to the Cisco WAE via Web cache communications protocol (WCCP) from the branch router. The Cisco WAE performs the following functions:
•Locally cached—If the data that is being requested is locally cached, the Cisco WAE responds to the requestor with the cached data and requests only required data from the server farm. This allows the WAN to become more efficient as only "needed data" requested.
•New data—If the data that is being forwarded to the server farm or coming from the server farm, the Cisco WAE performs compression algorithms on the data allowing for the WAN to become more efficient.
The WAN simulator provides simulations of the following WAN links:
1. WAN Type 1 (Intracontinental or T1)
a. Bandwidth - 1.544 Mbps, ESF, B8ZS, Delay - 100 mS, Loss - drop one packet in every 1000 packets (0.1%)
2. WAN Type 2 (Intercontinental)
a. Bandwidth - 512 Kbps, ESF, B8ZS, Delay - 200 mS, Loss - drop one packet in every 500 packets (0.2%)
The data center (DC) follows the design guidelines found in the Data Center Infrastructure Design Guide 2.5, a Cisco Validated Design found at http://www.cisco.com/go/designzone. The design consists of a data center WAN router, core, aggregation, and access Ethernet switching, and the server farm where the application resides. In this document, the focus is on the DC WAN router, aggregation, and the server farm. The core Ethernet switching provides routing to and from the DC WAN router and the aggregation. The access Ethernet switching provides Layer 2 connectivity for the server farms to the aggregation.
The DC WAN router performs the same function as the branch WAN router by redirecting traffic to the DC Cisco WAE. The DC Cisco WAE performs the following:
•Locally cached—If the data that is being requested is locally cached, the Cisco WAE responds to the requestor with the cached data and requests only required data from the branch. This allows the WAN to become more efficient as only "needed data" is requested.
•New data—If the data that is being forwarded to the branch or coming from the branch, the Cisco WAE performs compression algorithms on the data allowing for the WAN to become more efficient.
Included in the data center is the Cisco WAAS central manager (CM), which runs on the Cisco WAE Appliance. The Cisco WAAS CM provides a centralized mechanism for configuring Cisco WAAS features and reporting and monitoring Cisco WAAS traffic. It can manage a topology containing thousands of Cisco WAE nodes and be accessed from any Web browser using SSL. The Cisco WAAS CM can be configured for high availability by deploying a pair of Cisco WAE appliances as central managers.
Within a Cisco WAAS topology, each Cisco WAE runs a process called central management system (CMS). The CMS process provides SSL-encrypted bidirectional configuration synchronization of the Cisco WAAS CM and the Cisco WAE appliances. The CMS process is also used to exchange reporting information and statistics at a configurable interval. When the administrator applies configuration or policy changes to a Cisco WAE appliance or a group of Cisco WAE appliances, the Cisco WAAS Central Manager automatically propagates the changes to each of the managed Cisco WAEs. Cisco WAEs that are not available to receive the changes will receive them the next time the appliances become available.
The aggregation segment contains Cisco ACE, which provides the following features:
•Virtualization—Virtualization is device partitioning into multiple contexts, where each context can be configured for different applications and is independent of any others. In the Joint Solution, Cisco ACE is configured with the Admin context and the SharePoint context. Note that the Cisco ACE can support up to 250 contexts.
•Session persistence—Session persistence is the ability to forward client requests to the same server for the duration of the session. MOSS requires either source Internet Protocol (IP) based session persistence or Hypertext Transfer Protocol (HTTP) cookie based session persistence.
•Transparent interception—Transparent interception performs a Network Address Translation (NAT) function to conceal the real server IP address that is residing in the server farm. The SharePoint context is configured with a Virtual IP (VIP) that provides a single address that users use to connect to the server farm. This allows users to access the MOSS application by placing a single IP in the Web browser.
•Allowed server connections—Allowed server connections is the maximum number of active connections value on a per-server basis and/or globally to the server farm.
•Health monitoring—Health monitoring is used to track the state of the server and determine its ability to process connections in the server farm. The SharePoint context used a compound probe to determine if servers are operational and responding to HTTP requests.
Cisco ACE provides load balancing of the traffic to the server farm using one of the following methods: Round Robin, Weighted Round Robin, Least Connections, Hash address, Hash cookie, Hash Header, and Hash URL. In the Joint Solution, Least Connections was used, which selects the server with the fewest number of server connections. Cisco ACE is also used to provide SSL offload and TCP reuse.
Inter-chassis Cisco ACE redundancy was used, in which a Cisco ACE module in one Cisco Catalyst 6500 Series Switch chassis is protected by a Cisco ACE module in a peer Cisco Catalyst 6500 Series Switch chassis connected by a fault tolerant (FT) VLAN. The FT VLAN is used to transmit flow-state information, configuration synchronization information, and the redundancy heartbeat.
The server farm consists of four physical servers—two application servers, one database server, and one active directory server
Two physical application servers, load balanced by Cisco ACE, each have the following hardware and software configuration:
•Microsoft Office SharePoint Server (MOSS) 2007
•Microsoft Windows Server 2003 Enterprise Edition Service Pack 1 including Windows SharePoint Server (WSS) and Internet Information Services
•Two dual core Intel Xeon process - 3.06 GHz
•1 GB RAM
•Two mirrored 36.4 GB serial ATA hard drives
The database server, which is connected to both application servers, has the following hardware and software configuration:
•Microsoft SQL Server 2005
•Microsoft Windows Server 2003 Enterprise Edition Service Pack 1
•Two dual core Intel Xeon process - 3.4 GHz
•3.5 GB RAM
•Two mirrored 72.8 GB serial ATA hard drives
The active directory server has the following hardware and software configuration:
•Microsoft Windows Server 2003 Enterprise Edition Service Pack 1
•Two dual core Intel Pentium 3 - 697MHz GHz
•256 MB RAM
•14 GB SATA hard drive
The database and active directory servers are not load balanced. No virtualization was used on any of the servers.
Figure 4 illustrates the SharePoint solution server deployment details in the data center.
Figure 4 SharePoint Solution Server Deployment Details in the Data Center
Packet Flow Without Cisco WAAS and Cisco ACE
Application packet flow from a remote site can be categorized into three segments, client, WAN, and server.
Figure 5 Normal Packet Flow
The client segment is defined as the location into which users are connected that allows them to obtain or retrieve data from the application residing on the server farm. Users have connected personal computers (PC) to a local external switch or an integrated switch/router. When a user opens a browser and provides a URL that points to the application residing on the server, the data is sent from the PC to the switch. The switch forwards the data to the router that connects to the wide area network (WAN).
The WAN provides the connectivity from the client location to the data center where the server farm is located. The WAN is provided by a service provider (SP) with a given service level agreement (SLA). The WAN inherently introduces delay and packet loss to the data traffic (packets).
The server segment consists of a highly available and resilient core, aggregation, and access Ethernet switching. The core routes the data traffic to and from the WAN and the aggregation layer. The aggregation layer provides consolidation of multiple access layers and routes the access layer traffic into the core. The aggregation layer also takes the data traffic from the core layer and sends it to the appropriate access layer. The access layer provides connectivity to the server farm where the applications reside. The data traffic (URL, per the example) from the client segment transverses the data center until the data traffic is received by the appropriate server. The server's application responds to the request and responds back to the user by forwarding the appropriate data back the client segment.
Transaction response time consists of server response time and WAN round trip time. Delays in the WAN or the time to process a request on a server lead to a longer wait times for data to be viewed by the end user.
Packet Flow with Cisco WAAS and Cisco ACE
Figure 6 Packet Flow with Cisco WAAS and Cisco ACE
The following sequence describes the handshake between a client and the server farm and the data transfer phase:
1. The client sends a TCP SYN (synchronize) packet to the server farm VIP address. The packet is forwarded to the branch router. The branch router intercepts the packet with WCCP and forwards it to the branch Cisco WAE appliance.
2. a.) The branch Cisco WAE applies a new TCP option (0x21) to the packet if the application is identified for optimization by an application classifier. The branch Cisco WAE adds its device ID and application policy support to the new TCP option field. This option is examined and understood by other Cisco WAEs in the path as the ID and policy fields of the initial Cisco WAE device. The initial ID and policy fields are not altered by another Cisco WAE. The packet is forwarded to the branch router and then to the WAN. b.) During the data transfer phase, if the requested data are in its cache, the branch Cisco WAE returns its cached data to the client. Traffic does not travel through the WAN to the server farm. Hence both response time and WAN link utilization are improved.
3. The packet arrives on the WAN edge router. The WAN edge router intercepts the packet with WCCP and forwards the packet to the data center Cisco WAE.
4. The data center Cisco WAE inspects the packet. Finding that the first device ID and policy is populated, it updates the last device ID field (first device ID and policy parameters are unchanged). The data center Cisco WAE forwards the packet to the WAN edge router. The edge router forwards it to the Cisco ACE. The Cisco ACE forwards the packet to the server farm VLAN with TCP option 21 removed. TCP options are usually ignored by the server, even if it is still in place. The Cisco ACE performs load balancing to the data traffic. Other functions the Cisco ACE performs include SSL offload, TCP reuse, cookie and IP sticky pertinence.
5. The following steps are for reverse traffic flow. The server farm sends the SYN/ACK packet back to the client with no TCP option. The packet from the server farm VLAN is matched and forwarded to the Cisco ACE and then to the WAN edge router. The WAN edge router forwards the packet to the data center Cisco WAE. The data center Cisco WAE marks the packet with TCP option 0x21. During the data transfer phase, the data center Cisco WAE caches the data if the data are not in its cache.
6. The data center Cisco WAE sends the packet to the WAN edge router.
7. The packet travels through the WAN and arrives at the branch router. The branch router intercepts the packet and forwards it to the branch Cisco WAE. The branch Cisco WAE is aware of the Cisco WAE in the data center because the SYN/ACK TCP option 0x21 contains an ID and application policy. The auto-negotiation of the policy occurs as the branch Cisco WAE compares its application-specific policy to that of its remote peer defined in the TCP option. At this point, the data center Cisco WAE and branch Cisco WAE have determined the application optimizations to apply on this specific TCP flow. During the data transfer phase, the branch Cisco WAE caches the data if the data are not in its cache.
8. The packet is forwarded to the branch router and then to the client.
Implementing and Configuring the Cisco ACE Solution
The Cisco ACE module used in this solution is deployed in a Cisco Catalyst 6509 switch in the data center aggregation layer. The Cisco ACE module is deployed in routed mode where the client and server side VLANs each support unique IP subnets. In this deployment mode the Cisco ACE acts as the default gateway for the application servers.
What Was Implemented
Key features implemented on the Cisco ACE module to support this application are:
•Layer 4 load balancing
•Persistence based on source IP address
•Layer 7 load balancing for SSL termination and TCP reuse
•Server health monitoring
•Connection replication for stateful failover
What Was Not Implemented/Tested
The following was not implemented in this solution:
•Persistence based on HTTP cookie
Figure 7 Network Topology
Hardware or Components
Table 1 Hardware
Product Chassis Modules Interfaces Memory
Must be inserted into a Cisco Catalyst 65XX chassis
Note For the data center infrastructure, refer to the Data Center Design and Implementation Guide at http://www.cisco.com/go/designzone.
Table 2 Software
Product Software/Code Version
Features and Functionality
Features, Services, and Application Design Considerations
Cisco ACE module support three primary types of session persistence, source IP based, HTTP cookie inserted by the application server, and HTTP cookie inserted by the Cisco ACE module. MOSS can work with each of these types of session persistence. Considerations used to determine which is best are based on deployment environment. If MOSS is deployed in an intranet environment, then simple source IP based sticky is probably adequate. If MOSS is deployed facing the Internet then a the problem of users coming through mega-proxy servers becomes an issue and it is probably better to use HTTP based cookie session persistence. The cookie contains information that the Cisco ACE uses to ensure persistence to a specific real server. Refer to Configuration Task Lists and Appendix A—Cisco ACE Configuration for configuration information.
High Availability, Scalability, and Redundancy
The Cisco ACE also offers multi-tiered redundancy, availability, and scalability for maximum protection. It is the only product in the industry that offers three types of high availability:
•Inter-chassis—A Cisco ACE in one Cisco Catalyst 6500 is protected by a Cisco ACE in a peer Cisco Catalyst 6500.
•Intra-chassis—A Cisco ACE in a Cisco Catalyst 6500 is protected by another Cisco ACE in the same Cisco Catalyst 6500 (the Cisco Catalyst 6500 has solid redundancy built in).
•Inter-partition—Cisco ACE supports high availability between virtual partitions configured across two modules to allow specific partitions to fail over without affecting the other partitions and applications on a given module.
All these application availability modes provide rapid, stateful application redundancy with replication of connection state and sticky tables.
Configuration Task Lists
This section describes the steps necessary to configure the equipment.
Installing Cisco ACE and MSFC Configuration
A Cisco ACE module interacts with clients and servers via VLANs that are set up in the Cisco Catalyst 6500 Series/Cisco 7600 Series Supervisor Engine 720 (Sup720). These VLANs must be configured on the Sup720 to be allowed to be sent to the Cisco ACE module. Without this configuration, by default Cisco ACE does not receive traffic from any VLAN.
The following sample configuration steps are performed on the MSFC. Refer to Appendix A—Cisco ACE Configuration for a complete configuration.
Step 1 Add Cisco ACE VLANs and database server VLAN. For example:vlan 210name ACE-CLIENT!vlan 211name ACE-SERVER!vlan 500name ACE-FT-VLAN!
Step 2 Add the SVCLC configuration.
For this deployment, Cisco ACE is installed in slot 3 in the Cisco Catalyst 6500 chassis. The following configuration needs to be added to allow Cisco ACE-specific VLAN traffic to be directed towards Cisco ACE:svclc multiple-vlan-interfacessvclc module 3 vlan-group 1svclc vlan-group 1 210,211,500
Step 3 Add the Switch Virtual Interface (SVI) configuration.
The SVI (interface VLAN) configuration defines Layer 3 instance on the router MSFC. The Cisco ACE client side VLAN SVI configuration is:interface Vlan210description ACE Client Side VLANip address 10.1.210.2 255.255.255.0standby 210 ip 10.1.210.1standby 210 Priority 120
Virtualization is a method to allocate available Cisco ACE resources into two or more contexts for security and management purposes. Up to 250 (5 with no additional license requirements) contexts can be configured on Cisco ACE. Resources can be allocated to each context to avoid a single context consuming the entire pool of resources. This document only covers key virtualization configuration. Within each context, Domains and Role Base Access Controls (RBACs) can be further configured to provide additional security and access control to the resources.
Sample context configuration steps are:
Step 1 Configure resource-class(es):ACE_1/Admin(config)# resource-class Gold<cr> Carriage return.
The different resources that can be segmented are:ACE_1/Admin(config-resource)# limit-resource ?acl-memory Limit ACL memoryall Limit all resource parametersbuffer Set resource-limit for buffersconc-connections Limit concurrent connections (thru-the-box traffic)mgmt-connections Limit management connections (to-the-box traffic)proxy-connections Limit proxy connectionsrate Set resource-limit as a rate (number per second)regexp Limit amount of regular expression memorysticky Limit number of sticky entriesxlates Limit number of Xlate entries
Step 2 Configure context(s):
A context is configured by giving it a name, allocating VLANs, and assigning it to a resource-class (previous step):context SharePointdescription SharePoint Testingallocate-interface vlan 210-211member Gold
Step 3 To configure per-context features and functionality, simply change to the context created in step 2. At that point, you have accessed a virtually new Cisco ACE context.ACE_1/Admin# changeto SharePoint
To provide high availability and redundancy, Cisco ACE can be set up and configured in a redundant mode. Cisco ACE can be configured in a typical active/backup redundancy mode or active/active (per context) redundancy mode.! Configure FT interface!ft interface vlan 500ip address 10.1.500.1 255.255.255.248peer ip address 10.1.500.2 255.255.255.248no shutdown!! Configure FT peer!ft peer 1ft-interface vlan 500heartbeat count 3heartbeat interval 1000!! Configure interface tracking!ft track interface VLAN210priority 200!! Create a fault tolerant group!ft group 1peer 1priority 200preemptassociate-context admininservice!
By assigning context(s) to a FT group, a network administrator can create multiple groups for multiple contexts where the ACTIVE contexts can be distributed among the two Cisco ACE modules. This setup provides active/active redundancy setup for load sharing and high availability.
Remote Management Access
To access the Cisco ACE module remotely using Telnet, SSH, SNMP, HTTP, or HTTPS or to allow ICMP access to the Cisco ACE module, a policy must be defined and applied to the interface(s) the access is entering.
The configuration steps in this section are required for both the Admin context and the application context. The following example is for the application context. Refer to Appendix A—Cisco ACE Configuration for a complete configuration.
Step 1 Configure class-map of type management:class-map type management match-any REMOTE-MGMT10 match protocol ssh any20 match protocol telnet any30 match protocol icmp any40 match protocol http any50 match protocol https any
Step 2 Configure policy-map of type management:policy-map type management first-match REMOTE-ACCESSclass REMOTE-MGMTpermit
Step 3 Apply policy-map to the VLAN interfaces:interface vlan 210service-policy input REMOTE-ACCESSinterface vlan 211service-policy input REMOTE-ACCESS
Configuring Interface(s) and Default Gateway
Interface VLANs need to be configured for Layer 3 connectivity to Cisco ACE. Service policies for load balancing, security, and management access to Cisco ACE are also applied at the interface VLAN level.
The configuration steps in this section are required for both the Admin context and the application context. The following example is for the application context. Refer to Appendix A—Cisco ACE Configuration for a complete configuration.
Step 1 Define an access-list to permit/deny traffic through Cisco ACE. For example:access-list ANYONE line 10 extended permit icmp any anyaccess-list ANYONE line 20 extended permit ip any any
Step 2 Configure IP address and network mask of the interface(s):interface vlan 220ip address 10.1.210.5 255.255.255.0peer ip address 10.1.210.6 255.255.255.0alias 10.1.210.4 255.255.255.0interface vlan 211ip address 10.1.211.2 255.255.255.0peer ip address 10.1.211.3 255.255.255.0alias 10.1.211.1 255.255.255.0
Step 3 Apply management access policy and access-group to the interface(s), no shut of the interface(s):interface vlan 210access-group input ANYONEaccess-group output ANYONEservice-policy input REMOTE-ACCESSno shutdowninterface vlan 211access-group input ANYONEaccess-group output ANYONEservice-policy input REMOTE-ACCESSno shutdown
Step 4 Default gateway can be configured as:ip route 0.0.0.0 0.0.0.0 10.1.210.1
Step 5 Verify interfaces are recognized by MSFC and operational.
Type show interface and verify the VLANs are up and assigned from the supervisor.
Here is an example of a working output:vlan210 is upHardware type is VLANMAC address is 00:1b:d5:9b:88:edVirtual MAC address is 00:0b:fc:fe:1b:02Mode : routedIP address is 10.1.210.5 netmask is 255.255.255.0FT status is activeDescription:Client side vlanMTU: 1500 bytesLast cleared: neverAlias IP address is 10.1.210.4 netmask is 255.255.255.0Peer IP address is 10.1.210.6 Peer IP netmask is 255.255.255.0Assigned from the Supervisor, up on Supervisor53808467 unicast packets input, 17900167965 bytes7331701 multicast, 7776 broadcast0 input errors, 0 unknown, 0 ignored, 0 unicast RPF drops91028995 unicast packets output, 5455629020 bytes4 multicast, 5202 broadcast0 output errors, 0 ignoredvlan211 is upHardware type is VLANMAC address is 00:1b:d5:9b:88:edVirtual MAC address is 00:0b:fc:fe:1b:02Mode : routedIP address is 10.1.211.2 netmask is 255.255.255.0FT status is activeDescription:Server side vlanMTU: 1500 bytesLast cleared: neverAlias IP address is 10.1.211.1 netmask is 255.255.255.0Peer IP address is 10.1.211.3 Peer IP netmask is 255.255.255.0Assigned from the Supervisor, up on Supervisor83222640 unicast packets input, 95861661879 bytes1118208 multicast, 47974 broadcast0 input errors, 0 unknown, 0 ignored, 0 unicast RPF drops53089290 unicast packets output, 4304456323 bytes4 multicast, 14950 broadcast0 output errors, 0 ignored
Cisco ACE uses probe, one of the available keep-alive methods, to verify the availability of a real server. Probe is configured by defining its type and name.
There are different types of probes that can be configured on Cisco ACE:ACE_1/Admin(config)# probe ?dns Configure dns probeecho Configure echo probefinger Configure finger probeftp Configure ftp probehttp Configure http probehttps Configure https probeicmp Configure icmp probeimap Configure imap probeldap Configure ldap probepop Configure pop proberadius Configure radius probescripted Configure script probesmtp Configure smtp probetcp Configure tcp probetelnet Configure telnet probeudp Configure udp probe
Some key timers and parameters need to be tuned when probes are configured. The value for these parameters influences how rapidly Cisco ACE (or any load balancer) takes a server out of rotation and brings it back in service.
The following parameters need to be tuned for probes of any type (icmp, udp, tcp, http, https, scripted):
•faildetect—Refers to how many consecutive failed probes qualify a server to declared probe failed. faildetect is configured as a counter value. The default value is 3. Generally, the faildetect value is left at the default value.
•interval—Refers to how frequently Cisco ACE sends probe to a server. The interval is configured in seconds. The default value is 120 seconds. Generally, the interval is configured around 5-10 seconds depending upon the applications and size of the environment.
•passdetect—Determines how Cisco ACE re-probes the server after it has been declared failed. The passdetect variable has two attributes:
–passdetect count—Refers to how many consecutive successful responses Cisco ACE needs to see before declaring a server as operational. The default value is 3. This value can be tuned according to the requirements.
–passdetect interval—Refers to how many seconds Cisco ACE waits to probe a server after it has been declared failed. The default value is 300 seconds. Generally, the value is changed to a much lower value in the 15-30 seconds range.
These additional parameters should be configured for TCP, HTTP, and HTTPS probes:
•Open—Refers to the time (in seconds) that Cisco ACE waits to keep a TCP connection open. The default value is 10 seconds. Generally this value is configured close to the interval value.
•Receive—Once a TCP SYN (for a probe) is sent to a server, the value for receive determines how long Cisco ACE waits to receive a reply from the server. This value is configured in seconds and the default value is 10 seconds. Generally it is configured as equal to or less than the value interval.
•Connection—This parameter determines how Cisco ACE closes the connection after it has successfully sent a probe. By default, Cisco ACE closes the connection gracefully, which means it sends TCP FIN to close the connection. Optionally, Cisco ACE can be configured to close the connection with a TCP RESET by configuring connection term forced.
•Port—TCP/UDP port number on which this probe is sent. The default values for various probes are:
•Request—Used to configure the HTTP Request method (HEAD or GET) and URL for the probe. The default method is GET and default URL is /. Generally, method and URL are configured according to specific applications.
This parameter is only applicable to HTTP/HTTPS probes.
•Expect—Allows Cisco ACE to detect two values from the server:
–expect status—The HTTP status code (or range) to expect from the server. There is no default HTTP return code expected; it has to be explicitly configured.
–expect regex—A regex can be configured to parse a specific field in the response data.
This parameter is only applicable to HTTP/HTTPS probes.
•SSL—Configured to define what cipher and SSL version Cisco ACE should use when sending an HTTPS probe. Ciphers and SSL versions supported on Cisco ACE are:ssl cipher:RSA_EXPORT1024_WITH_DES_CBC_SHA EXP1024-DES-CBC-SHA CipherRSA_EXPORT1024_WITH_RC4_56_MD5 EXP1024-RC4-MD5 CipherRSA_EXPORT1024_WITH_RC4_56_SHA EXP1024-RC4-SHA CipherRSA_EXPORT_WITH_DES40_CBC_SHA EXP-DES-CBC-SHA CipherRSA_EXPORT_WITH_RC4_40_MD5 EXP-RC4-MD5 CipherRSA_WITH_3DES_EDE_CBC_SHA 3DES-EDE-CBC-SHA CipherRSA_WITH_AES_128_CBC_SHA AES-128-CBC-SHA CipherRSA_WITH_AES_256_CBC_SHA AES-256-CBC-SHA CipherRSA_WITH_DES_CBC_SHA DES-CBC-SHA CipherRSA_WITH_RC4_128_MD5 RC4-MD5 CipherRSA_WITH_RC4_128_SHA RC4-SHA Cipherssl versions:SSLv2 SSL Version 2.0SSLv3 SSL Version 3.0TLSv1 TLS Version 1.0
This parameter is only applicable to HTTPS probes.
In the ANS solution for MOSS 2007, the Cisco ACE is configured to perform a compound health check to first determine if the server is up and running and then to determine if Microsoft Internet Information Server (IIS) is running. This is done by creating an ICMP probe to the server and then TCP port 80 probe.
The following are configurations for ICMP and HTTP:
•ICMP probe:probe icmp PINGinterval 2faildetect 2
•HTTP probe:probe tcp HTTPport 80interval 2faildetect 2passdetect interval 10passdetect count 2
Cisco ACE selects the real servers (rserver) to send the intended traffic based on certain sets of criteria. When configuring an rserver, be aware that an rserver name is case sensitive. The minimum configuration needed for rserver is the IP address and stating the rserver as inservice.
The same rserver can be used in multiple server farms (shown later in the document). If an rserver is made "no inservice" at the rserver level, then it is taken out of rotation from every server farm on which it is configured. This provides the flexibility to take a server completely out of rotation with a single command.
To take a server out of rotation on a per-server farm basis, rserver should be made "no inservice" at the server farm level.
The following is an example of configuring rserver on Cisco ACE:rserver host SP1ip address 10.1.210.10inservice
A server farm is a logical collection of real servers (rservers) that Cisco ACE selects based on certain sets of criteria. As with real server, serverfarm name is also case sensitive.
Basic server farm configuration includes adding rservers and probes to the server farm.
Key configuration options within server farm sub-configuration mode are:
•failaction—Defines what action Cisco ACE should take about currently established connections if a real is detected as probe_failed. The default behavior for Cisco ACE is to take no action and allow the connections to close gracefully or timeout.
A configurable option is failaction purge, which forces Cisco ACE to remove the connections established to that real and send TCP RST(s) towards the client(s) and real(s).
•predictor—Refers to the load balancing algorithm for the server farm. Options are:
–hash—Based on source/destination IP address, URL, cookie, and header
–leastconns—Based on least number of connections. By default slow start is enabled for leastconns and its timing can be tuned using predictor leastconns slowstart.
<1-65535> Specify slowstart duration in seconds
–roundrobin—Load balance in a roundrobin fashion (default).
•probe—Allows a probe to be applied to the server farm. Multiple probes can be applied to the same server farm.
•retcode—Used to configure server health-checks based on the HTTP return code. The configuration allows you to define a range of HTTP return codes and take an action once a threshold is reached.retcode <min> <max> check <remove|count|log> <threshold value> resume-service <value in seconds>
•rserver—Used to associate real server(s) with a server farm. Port address translation, maximum and minimum connections, and weight are some common configurations that can be done in rserver sub-configuration mode.
•transparent—When configured, Cisco ACE does not NAT Layer 3 IP address from VIP to real server's IP address.
The following is an example of basic server farm configuration:serverfarm host MOSSpredictor leastconnsprobe PINGprobe HTTPrserver SP1inservicerserver SP2inservice
Layer 4 Load Balancing
Cisco ACE uses class-map, policy-map, and service-policy to classify, enforce, and take action on incoming traffic. Traffic trying to reach a virtual IP address on a certain port can be used as classifiers for Layer 4 load balancing as the classification is only based on destination IP and destination port.
Note In the Joint Solution, Layer 4 load balancing was used to demonstrate basic load balancing and Layer 7 load balancing was used to demonstrate SSL termination and TCP reuse. The details are discussed in Layer 7 Load Balancing.
The following example shows the configuration steps needed:
Step 1 Configure virtual IP address (VIP) using class-map of type match-all:class-map match-all VIP-HTTP-102 match virtual-address 10.1.210.10 tcp eq 80
Step 2 Configure policy-map of type loadbalance to associate sticky server farm:policy-map type loadbalance first-match VIP-POLICY-10class class-defaultsticky-serverfarm SRC-IP-STICKY
Step 3 Configure policy-map of type multi-match to associate class-map configured in step 1 above. Also apply ssl-proxy server under class maps for HTTPS traffic:policy-map multi-match LB-VIPclass VIP-HTTP-10loadbalance vip inserviceloadbalance policy VIP-POLICY-10loadbalance vip icmp-reply
Step 4 Apply policy-map to the interface VLAN:interface vlan 210service-policy input LB-VIP
Layer 7 Load Balancing
Similar to Layer 4 policy, Cisco ACE uses class-map, policy-map, and service-policy to classify and enforce a Layer 7 policy. Cisco ACE uses additional information such as URL, HTTP header, or cookie to make a load balancing decision. The following example shows the configuration steps for URL-based matching. Similar steps can be used for cookie or header matching.
The following example shows the configuration steps needed:
Step 1 Configure class-map of type HTTP:class-map type http loadbalance match-any L7-URL2 match http url .*.*
Step 2 Configure HTTP parameters (optional):parameter-map type http L7-mapcase-insensitive
Step 3 Configure virtual IP address (VIP) using class-map of type match-all:class-map match-all VIP-HTTP-102 match virtual-address 10.1.210.10 tcp eq 80
Step 4 Configure policy-map of type loadbalance to associate server farm:policy-map type loadbalance first-match L7-matchclass L7-URLsticky-serverfarm STICKY-INSERT-COOKIEclass class-defaultserverfarm MOSS
Step 5 Configure policy-map of type multi-match to associate class-map configured in step 1 above. See SSL Termination for configuring SSL VIP.policy-map multi-match LB-VIPclass VIP-HTTP-10loadbalance vip inserviceloadbalance policy L7-matchloadbalance vip icmp-reply
Step 6 Apply policy-map to the interface VLAN:interface vlan 210service-policy input LB-VIP
Stickiness (Session Persistence)
Session persistence or sticky configuration allows multiple connections from the same client to be sent to the same real server by Cisco ACE. Cisco ACE supports stickiness based on source and destination IP address and HTTP cookies. Cisco ACE insert cookie persistence is when the Cisco ACE inserts the cookie on behalf of the server upon the return request, so that the Cisco ACE can perform cookie stickiness even when the servers are not configured to set cookies. The cookie contains information that the Cisco ACE uses to ensure persistence to a specific real server.
The following sample configuration is for source IP-based sticky.
Source IP Stickiness
Step 1 Configure a sticky group:sticky ip-netmask 255.255.255.255 address source SRC-IP-STICKYtimeout 10serverfarm MOSS
Step 2 Apply sticky group to a loadbalance policy as a sticky-serverfarm:policy-map type loadbalance first-match VIP-POLICY-10class class-defaultsticky-serverfarm SRC-IP-STICKY
Step 3 Apply load balance policy to a multimatch policy:policy-map multi-match LB-VIPclass VIP-HTTP-10loadbalance vip inserviceloadbalance policy L7-MATCHloadbalance vip icmp-reply
Step 4 Apply multimatch policy as a service-policy to the interface VLAN:interface vlan 210description Client side vlanip address 10.1.210.5 255.255.255.0alias 10.1.210.4 255.255.255.0peer ip address 10.1.210.6 255.255.255.0access-group input ANYONEservice-policy input LB-VIPservice-policy input REMOTE-MANAGEMENTno shutdown
SSL termination configuration on Cisco ACE provides SSL traffic termination on Cisco ACE instead of on the servers. This allows the offloading of server resources and also provides HTTP request inspection for various load balancing functionality.
Front End SSL Termination
With SSL termination on the Cisco ACE, client to Cisco ACE traffic is SSL encrypted, but Cisco ACE to server traffic is clear-text. The configuration steps to implement front end SSL termination are:
Step 1 Generate key:ACE_1/testfeature# crypto generate key 512 testkey.keyACE_1/testfeature# show crypto key allFilename Bit Size Type-------- -------- ----testkey.key 512 RSA
Step 2 Define CSR parameters set:crypto csr-params testparamscountry USstate Californialocality SJorganization-name CMOorganization-unit ANScommon-name www.testssl.comserial-number cisco123
Step 3 Generate csr:ACE_1/testfeature# crypto generate csr testparams testkey.key-----BEGIN CERTIFICATE REQUEST-----MIIBHjCByQIBADBkMQswCQYDVQQGEwJVUzETMBEGA1UECBMKQ2FsaWZvcm5pYTELMAkGA1UEBxMCU0oxCzAJBgNVBAoTAkFTMQwwCgYDVQQLEwNUQVMxGDAWBgNVBAMTD3d3dy50ZXN0c3NsLmNvbTBcMA0GCSqGSIb3DQEBAQUAA0sAMEgCQQC+xphqQJn9EOzOhkFfVCVO5SYJj7nVjWmaslVZOi7TYKzFgXtJexMt0Y1VyO7XY+U5XdZuvoxEcO4rdAGzo84HAgMBAAGgADANBgkqhkiG9w0BAQQFAANBAAL9EzKcYyOrL3XYc7YGSTgpa1B8tTpCpJIVwrHwolyK3OzvfudLTbF7CQ2V3jUYS//sf2Cei8fe+voKIQE9nI4=-----END CERTIFICATE REQUEST-----
Step 4 Obtain certificate:
An SSL certificate can be obtained from various Certificate Authority (CA) companies like VERISIGN. The following example shows using a CISCO router as a CA:CMO-CA-SERVER#crypto pki server ANS-CA request pkcs10 terminal pem% Enter Base64 encoded or PEM formatted PKCS10 enrollment request.% End with a blank line or "quit" on a line by itself.-----BEGIN CERTIFICATE REQUEST-----MIIBHjCByQIBADBkMQswCQYDVQQGEwJVUzETMBEGA1UECBMKQ2FsaWZvcm5pYTELMAkGA1UEBxMCU0oxCzAJBgNVBAoTAkFTMQwwCgYDVQQLEwNUQVMxGDAWBgNVBAMTD3d3dy50ZXN0c3NsLmNvbTBcMA0GCSqGSIb3DQEBAQUAA0sAMEgCQQC+xphqQJn9EOzOhkFfVCVO5SYJj7nVjWmaslVZOi7TYKzFgXtJexMt0Y1VyO7XY+U5XdZuvoxEcO4rdAGzo84HAgMBAAGgADANBgkqhkiG9w0BAQQFAANBAAL9EzKcYyOrL3XYc7YGSTgpa1B8tTpCpJIVwrHwolyK3OzvfudLTbF7CQ2V3jUYS//sf2Cei8fe+voKIQE9nI4=-----END CERTIFICATE REQUEST-----Quit% Granted certificate:-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----IOS-CA-SERVER#
Step 5 Import certificate on Cisco ACE:ACE_1/testfeature# crypto import terminal testcert.pemPlease enter PEM formatted data. End with "quit" on a new line.-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----quit
Step 6 Validate certificate using key:ACE_1/testfeature# crypto verify testkey.key testcert.pemKeypair in testkey.key matches certificate in testcert.pem.
Step 7 Configure SSL parameters and SSL proxy service:
a. SSL parameter configuration:parameter-map type ssl sslparamscipher RSA_WITH_RC4_128_MD5version SSL3
b. SSL proxy service configuration:ssl-proxy service testsslkey testkey.keycert testcert.pemssl advanced-options sslparams
Step 8 Configure class-map (for VIP) and policy-maps:serverfarm host MOSS-Farm2probe HTTPrserver SP1 80inservicerserver SP2 80inserviceclass-map match-all VIP-SSL-1752 match virtual-address 10.74.1.175 tcp eq httpspolicy-map type loadbalance first-match vip-ssl-175class class-defaultserverfarm MOSS-Farm2policy-map multi-match LB-VIPclass VIP-WEB-175loadbalance vip inserviceloadbalance policy L7-matchloadbalance vip icmp-replyappl-parameter http advanced-options L7-mapclass VIP-SSL-175loadbalance vip inserviceloadbalance policy vip-ssl-175loadbalance vip icmp-replyssl-proxy server testssl
Step 9 Apply multi-match policy-map to service-policy at interface level or globally:interface vlan 210ip address 10.1.210.5 255.255.255.0access-group input everyoneaccess-group output everyoneservice-policy input REMOTE-ACCESSservice-policy input LB-VIPno shutdown
Configuration and Menus
See Appendix A—Cisco ACE Configuration for the configuration used to support the Joint Solution.
These show commands can help troubleshoot issues with the configuration:
•show stats—Displays the statistical information relating to the operation of the Cisco ACE.
•show service-policy policy_name—Displays the statistics for service policies enabled globally within a context or on a specific interface.
•show serverfarm name detail—Displays the summary or detailed server-farm statistics.
•show rserver rserver_name detail—Displays the summary or detailed statistics for a named real server or for all real servers.
•show probe—Displays the probe information including script probes.
•show arp—Displays the current active IP address-to-MAC address mapping in the ARP table, statistics, or inspection or timeout configuration.
•show arp statistics—Displays the ARP statistics for all VLAN interfaces.
•show context—Verifies the auto-sync configuration of all contexts.
•show ft group status—Verifies FT status of all configured context in the Cisco ACE.
•show ft peer detail—Verifies the state of FT peering.
•show resource usage—Displays the resource usage for each context.
•show np NP_number—Displays the hardware information stored on the three network processors.
Configuration rollback allows the administrator to revert back to a previous configuration when the new configuration does not work.
Step 1 Create a configuration checkpoint:ACE_1/testfeature# checkpoint create name
Step 2 Rollback to the checkpoint defined in step 1:ACE_1/testfeature# show checkpoint allACE_1/testfeature# checkpoint rollback config-05-09-06
Results and Conclusions
Microsoft Office SharePoint Server 2007 can be effectively deployed behind a Cisco ACE load balancer with very few special considerations. Standard Layer 4 load balancing is sufficient for most deployments, but Layer 7 load balancing can be used to support SSL termination and TCP reuse if necessary. For SSL encrypted deployments, the Cisco ACE module can off load server CPU utilization significantly. In addition, when coupled with TCP reuse feature, the Cisco ACE module can further reduce server CPU utilization. Test results for these two deployment modes is shown below.
In this CPU utilization testing scenario, the PC running the HP Mercury LoadRunner was directly attached to the aggregation switch in the data center to generate 700 concurrent SSL-encrypted users connecting to a SharePoint page on the application server, through Cisco ACE to determine benefits of SSL off-load and TCP reuse. The Cisco ACE module performed Layer 4 load balancing to the MOSS servers for SSL termination on the server and then the Cisco ACE was reconfigured for Layer 7 load balancing to terminate SSL on the Cisco ACE module itself. An additional test was run with SSL termination on Cisco ACE with TCP reuse enabled to determine incremental benefits form this feature.
As shown in the test results below, CPU utilization was decreased 50 percent with the termination of SSL encrypted traffic on the Cisco ACE. When TCP reuse was added, CPU utilization dropped another 21 percent, for a total of 71 percent CPU utilization savings. In these tests, the two application servers were each running two dual core 3.06 GHz CPUs with 1 GB of RAM.
Figure 8 Directly Attached With Cisco ACE
Test Results —Server CPU Statistics
Figure 9 SSL Terminated on Server
Figure 10 SSL Termination on Cisco ACE Module
Figure 11 SSL Termination on Cisco ACE Module with TCP reuse
Figure 12 CPU Utilization Decreased as a Result of SSL Decryption and TCP Reuse on Cisco ACE
Implementing and Configuring the Cisco WAAS Solution
The Cisco WAAS solution requires a minimum of three Cisco Wide Area Application Engine (WAE) appliances to auto-discover and deliver applicable application optimizations. One Cisco WAE is placed in the enterprise data center and the other at the branch site. The enterprise data center Cisco WAE is placed on the WAN edge connected to the WAN router. The third Cisco WAE is used for the Central Manager. The architecture offloads the Cisco WAE device from the local branch router and leverages the available ports on a local switch. This design provides scalability and availability for the solution.
What Was Implemented
Cisco WAAS technology requires the efficient and predictable interception of application traffic to produce results. It is critical that the Cisco WAE device see the entire TCP conversation. At the WAN edge, Cisco routers support the following four methods of traffic interception:
•Policy-based routing (PBR)
•Web Cache Communications Protocol (WCCP) v2
•Service policy with Cisco ACE
WCCPv2 is the most common method used in the remote branch environment; therefore, WCCPv2 has been leveraged for this solution.
Note Cisco WAEs "out of box" have a standard set of application variables and ports that are defined for optimization. In this solution no changes need to be made to the standard default configuration of the Cisco WAEs.
What Was Not Implemented
The following was not implemented in this solution:
•Cisco WAAS Network Module in which Cisco WAAS is installed in an integrated services router, providing a comprehensive solution within a single platform. This architecture provides less scalability and should be considered for use with a branch with small number of users.
Figure 13 Network Topology
Hardware or Components
Table 4 Hardware
Product Chassis Modules Interfaces Memory
2 10/100/1000 Ethernet, serial port
4 Gbytes, 144 GB SCSI HD
2 10/100/1000 Ethernet, serial port
2 Gbytes, 144 GB SCSI HD
Table 5 Software
Product Software/Code Version
Cisco WAAS Enterprise License for 1 Cisco WAE Appliance
Cisco WAAS Enterprise License for 1 Cisco WAE Appliance
Features and Functionality
Table 6 Features and Functionality
Product Supported Features and Functionality Used in the Solution
Transport Flow Optimization (TFO)
Data Redundancy Elimination (DRE), LZ compression
Features, Services, and Application Design Considerations
Most multi-tiered applications support Web-based clients in addition to native application clients. Web-based clients use port 80 to communicate to the Web server. Applications in this test use port 80. In the context of Cisco WAAS, port 80 is accelerated by default; no further configuration in the Cisco WAE is necessary unless the application requires ports that are not part of the default application profile. For applications that use TCP ports that are not defined in the default application profile, you must define ports to the existing application profile or create a new application profile with the associated ports. With the recommended design of Cisco WAAS at the WAN edge, client data only traverse the Cisco WAEs once, at the ingress/egress of the data center. Further application communication between the Web servers, application servers, and database servers are within the data center and are not affected by Cisco WAAS.
Transport Flow Optimization (TFO), Data Redundancy Elimination (DRE), and Lempel-Ziv (LZ)-compression, the three key technologies of Cisco WAAS, are enabled by default. Each of these features and functionalities are described in Features and Functionality. The net results are reduced traffic and decreased latency across the WAN. Since Cisco WAAS deployments are transparent to the network and application, applications do not need to be aware of the added functionalities and continue to work as-is, but with decreased response time and increased traffic throughput and transactions.
Additional information on Cisco WAAS data center and branch designs are available at:
•Enterprise Data Center Wide Area Application Services (WAAS) Design Guide http://www.cisco.com/en/US/netsol/ns751/networking_solutions_sub_program_home.html
•Enterprise Branch Wide Area Application Services Design Guide (Version 1.1) http://www.cisco.com/en/US/netsol/ns816/networking_solutions_program_home.html
Scalability and Capacity Planning
Cisco WAE farms can scale up to 32 devices with WCCP and up to 16,000 with Cisco ACE load balancing. Cisco WAAS services scale linearly in a N+1 configuration. In addition to the Max Optimized TCP connections, the fan out ratio between the DC Cisco WAE and branch Cisco WAE have to be considered. The fan out ratio is determined by a number of factors, such as the number of Cisco WAEs in the branch offices, amount of network traffic, and number of TCP connections. A sizing tool is available internally that can help automate sizing decisions. NetFlow, NetQoS, and other network analysis tools can provide additional traffic flow information for increased accuracy in scalability and capacity planning.
Branch devices range from the NME-WAE-302 for very small offices to the 612-4GB or even higher models for bigger branch sites. WAE 7326 and up are designed for data center installations.
Cisco WAAS deployments are transparent to the application. The application client and server do not know Cisco WAAS is optimizing traffic flows. High availability is built into the WCCP interception. When WCCP is not active or if Cisco WAAS devices are not functioning, WCCP does not forward traffic to the Cisco WAEs, resulting in un-optimized traffic flow. This is the worse case scenario; traffic flow continues but is not optimized.
Device High Availability
The Cisco WAEs have many built-in high availability features. The disk subsystem is recommended to be configured with RAID 1 protection. RAID 1 is mandatory when two or more drives are installed in the Cisco WAE. With RAID 1, failure of the physical drive does not affect normal operations. Failed disks can be replaced during planned downtime. Multiple network interfaces are available. Standby interfaces can be configured for interface failover. A standby interface group guards against network interface failure on the Cisco WAE and switch. When connected to separate switches in active/standby mode, the standby interface protects the Cisco WAE from switch failure.
Cisco WAEs and the network provide additional high availability (HA) capabilities. Routers can be configured redundantly providing HSRP or GLBP services. Cisco WAEs can configured in a N+1 configuration. N+1 configuration provides scalability and availability. This design calls for N number of Cisco WAEs for a specific workload, then add a standby Cisco WAE. Since the workload always distributes evenly among the Cisco WAEs, the standby Cisco WAE is utilized, reducing overall workload. In the event that a Cisco WAE fails, the rest of Cisco WAEs can resume normal workload.
Configuration Task Lists
The following subsections describe the information required prior to configuration of the equipment.
Central Manager (CM) is the management component of Cisco WAAS. CM provides a GUI for configuration, monitoring, and management of multiple branch and data center Cisco WAEs. CM can scale to support thousands of Cisco WAE devices for large-scale deployments. The CM is necessary for making any configuration changes via the Web interface. In the event of CM failure, Cisco WAEs continue to function, however changes cannot be made using the Web pages on the CM until the CM comes back online.
Cisco WAEs need to connect to the CM on the initial setup. The registration process adds the Cisco WAE to the CM and initializes the local Cisco WAE data base. When disk encryption on the Cisco WAE is enabled, the CM must be available to distribute the encryption key in the event the Cisco WAE reboots.
Centralized reporting can be obtained from the CM. Individually, the Cisco WAEs provide basic statistics via the CLI and local device GUI. System-wide application statistics are available from the CM GUI. Detailed reports such as total traffic reduction, application mix, and pass-through traffic can be obtained from CM GUI.
The following example shows the configuration steps needed to configure CM.
Note At least one Cisco WAE must be the Central Manager. Adding backup Central Managers increases availability. Central Managers should be installed in the data center with other critical servers, not near the branch- or WAN-facing segments.
Step 1 Configure the device to be Central Manager. It is set to application-accelerator mode by default:device mode central-manager
Step 2 Configure the Central Manager IP address:interface GigabitEthernet 1/0ip address 10.1.21.2 255.255.255.0
Step 3 Set up the default gateway:ip default-gateway 10.1.21.1
Step 4 Set the primary interface. Cisco WAAS supports multiple network interfaces type, port channels, and standby interface. Cisco WAAS uses the primary interface for traffic interception and delivery. The primary interface must defined:primary-interface GigabitEthernet 1/0
Step 5 Define the NTP server. Traffic statistics are captured and forward to Central Manager and NetQos. The time stamp on each packet needs to be accurate. All Cisco WAEs and routers should synchronize to the same NTP server:ntp server 10.1.6.20
Step 6 Initialize the Configuration Management System (CMS) database. The CMS contains configuration rules and information. The Central Manager is the repository of CMS data:cms enable
Step 7 Login to the CM Web GUI on port 8443 after the CM is up and running. The initial CM screen is an overview of the health of the system. It contains information on number of devices, status, application traffic, and optimization rate.
Figure 14 Cisco WAAS Central Manager
Branch and Data Center Router
The branch and data center router provides WCCP interception points for Cisco WAAS. Without WCCP interception, Cisco WAAS does not know where to obtain and optimize traffic flow. Different methods of interception and redirection are supported by routers and switches. Redirection methods depend on the speed requirement and router/switch platform. In this deployment, Generic Router Encapsulation (GRE) redirection is used.
The loopback interface on the router is essential for identifying the router ID. While any IP address can be used to identify the router ID, the loopback interface is preferred over the physical interfaces. Loopback interfaces are always available; there are no physical ties to them. Other routing protocols also use loopback interfaces as a preferred method for naming the router ID. With the IP address tied to a specific physical interface, when the physical interface goes down, the IP address becoming unavailable, causing unexpected issues with WCCP groups.
Step 1 Configure the loopback interface:interface Loopback0ip address 10.1.6.21 255.255.255.255
WCCP service 61 and 62 direct the router to re-route traffic from the interface to the WCCP group. Service 61 redirects ingress traffic and service 62 redirects egress traffic. Both service 61 and 62 are needed to complete redirect bi-directional traffic flow. WCCP is an open standard. Other equipment that implements the WCCP protocol can participate in the WCCP group. Passwords should be assigned to WCCP groups to prevent rogue traffic interception and redirection.
Step 2 Configure WCCP service 61 and 62 with a password:ip wccp 61 password ANSip wccp 62 password ANS
Step 3 Configure the Cisco WAE VLAN. The Cisco WAE needs to reside in its own subnet for WCCP interception:interface Vlan301description WAE vlan - 301ip address 10.1.12.1 255.255.255.0
Step 4 Exclude the WAW subnet from interception since we are using a single interface to intercept incoming and outgoing packets. The interception exclusion is required because the router does not discriminate traffic from the Cisco WAE for client/server. Traffic must be redirected to the Cisco WAE after it is optimized by the Cisco WAE; the effect would be forwarding loop.ip wccp redirect exclude in
Step 5 Enable the NetFlow collection for outgoing traffic from the Cisco WAEs:ip flow egress
Step 6 Assign the Cisco WAE VLAN to physical port:interface FastEthernet1/0description WAE portswitchport access vlan 301
Step 7 Configure the client VLAN. This is the VLAN or interface for WCCP interception:interface Vlan300description client vlan - 300ip address 10.1.11.1 255.255.255.0
Step 8 Configure WCCP interception service 61 and 62 on the client VLAN. All ingress/egress packets from this VLAN/interface are forwarded to the Cisco WAE for optimization:ip wccp 61 redirect inip wccp 62 redirect out
Step 9 NetFlow statistics is configured for all outbound traffic:ip flow egress
Step 10 Configure NTP to synchronize to a master clock. Traffic statistics are captured and forwarded to Central Manager and NetQos. The time stamp on each packet needs to be accurate. All Cisco WAEs and routers should synchronize to the same NTP server.ntp server 10.1.6.20
Step 11 Configure NetFlow to send information to the collector. Notice NetFlow also uses loopback interface as the source address. NetFlow sends statistics from the Cisco WAE and router to the NetFlow aggregator. NetFlow statistics can be overwhelming for smaller connections. It is recommended that Cisco WAAS optimize NetFlow transfers.ip flow-export source Loopback0ip flow-export version 5ip flow-export destination 10.1.70.10 9995
Step 1 Set device mode to accelerator. Cisco WAE can be set up as application accelerator or Central Manager. By default application-accelerator is enabled:device mode application-accelerator
Step 2 Configure the Cisco WAE IP addresses:interface GigabitEthernet 1/0ip address 10.1.20.2 255.255.255.0
Step 3 Set up the default gateway:ip default-gateway 10.1.20.2
Step 4 Set up the primary interface. Cisco WAAS supports many type of interfaces, including local network failover. Designating a primary interface is required. Cisco WAAS uses this interface for interception and redirection.primary-interface GigabitEthernet 1/0
Step 5 Turn on WCCP version 2:wccp version 2
Step 6 Add the router to the router list:wccp router-list 1 10.1.20.1
Step 7 Set up tcp promiscuous mode to accept all traffic from the interface. The WCCP password is the same for all devices in the WCCP group, including routers:wccp tcp-promiscuous router-list-num 1 password cisco
Step 8 Set up NTP server. Traffic statistics are captured and forwarded to Central Manager and NetQos. The time stamp on each packet needs to be accurate. All Cisco WAEs and routers should synchronize to the same NTP server.ntp server 10.1.20.1
Step 9 Set up Central Manager address. The Cisco WAE needs register to the Central Manager for statistics reporting and management. Configurations on a per device basis can be perform by the CLI and device GUI. Site wide or Cisco WAAS group have to be perform by the Central Manager. The Central Manager can run operations on thousands of Cisco WAEs at once, saving precious time managing the Cisco WAAS infrastructure:central-manager address 10.1.21.2
Step 10 Enable CMS. This command initializes the local database and connects to the Central Manager:cms enable
Step 11 Set up NetFlow to send Cisco WAAS statistics to the NetFlow Aggregator. Notice the host IP address is not the NetFlow Aggregator, but the management station. The management station opens another connection to the Cisco WAE to inform the IP address of the aggregator:flow monitor tcpstat-v1 host 10.1.70.11flow monitor tcpstat-v1 enable
Configuration and Menus
See Appendix B—Cisco WAE Configurations for the configuration used by MOSS.
Cisco WAE Commands
These show commands can help troubleshoot issues with the configuration:
•sh wccp status—Verifies WCCP V2 is enabled.
•sh wccp services—Verifies WCCP service 61 and 62 are active. Service 61 and 62 must be active.
•sh wccp routers—Verifies the router can see the Cisco WAE. Notice that the router ID is the router loopback address. Sent To is the router interface on the Cisco WAE VLAN. All routers are defined and visible on the Cisco WAE.
•sh tfo connection summary—Verifies Cisco WAAS clients are using Cisco WAAS for connectivity. Show tfo connections show all optimize path in the Cisco WAE. The policy field indicates which optimization method is active for the specified link. F shows the link is fully optimized, which includes DRE, TFO (shown as TCP Optimization), and LZ compression. Pass-through connections are connections that are not optimized.
•sh statistics dre—Checks DRE usage. There are two sections in the statistics. One is encode, traffic coming in to the Cisco WAE from the client/server. The Cisco WAE needs to compress the incoming traffic with LZ compression then apply DRE. Another is the decode, traffic coming from the peering Cisco WAE, DRE lookup is performed and traffic uncompressed. These statistics are useful for finding compressibility of the data.
•sh ip wccp 61—Verifies WCCP service 61 and 62 are active. This command shows global wccp information and how the packets are redirected. Redirect and group access-list issues are easier to troubleshoot with this output. Service 62 should also check with sh ip wccp 62.
•sh ip wccp 61 detail—Checks WCCP client hash or Layer 2 assignments. This command also checks the status of the WCCP client, namely the Cisco WAEs. sh ip wccp 61 shows global WCCP information; this command shows detailed WCCP client information. Hashing assignments (Cisco WAE bucket assignments), client ID, and client status are found in the output.
•sh ip wccp interface detail—Verifies which interface has WCCP configured. Identify all interfaces within a router or switch that have WCCP configured with ingress or egress for exclude-in redirection. Another way to get this information is with sh run; examine each interface.
•sh ip wccp 61 view—Verifies WCCP group membership. Check service 62 as well.
Results and Conclusions
Figure 15 Data Volume Chart 1—With and Without Cisco WAAS
Figure 15 shows the amount data volume traversing the 1.544 Mbps WAN link with and without the Cisco WAAS device that was observed during in a 30 minute cycle with 40 users performing site navigation on the MOSS application. Figure 15 is based on the total amount data passing through the WAN divided by the total number of transactions under the conditions identified earlier.
The Cisco WAAS device reduces the amount of unnecessary data that traverses the WAN by locally caching data and using compression algorithms on the data the must traverse the WAN. As seen in Figure 15, the Cisco WAAS-enabled network becomes more efficient as less data must traverse the WAN.
With this efficiency, the end-user transaction times are faster and more transactions can occur, as shown in the figures below. Figure 16 and Figure 17 indicate the number of transactions that were observed for the same 30 minute cycle.
Figure 16 shows the decreased average responses time with Cisco WAAS enabled. This would result in the end-user obtaining quicker responses from the application and allowing more transactions to occur as the end-user is not waiting as long for the responses, as seen in the transaction summary charts.
Figure 16 Transaction Times Chart
Figure 17 Transaction Summary Chart
Figure 18 Bandwidth Savings for 512 Kbps WAN Link
Figure 18 provides the amount data volume traversing the 512 Kbps WAN link with and without the Cisco WAAS device that was observed during in a 30 minute cycle with 40 users performing site navigation on SharePoint. The chart is based on the total amount data passing through the WAN divided by the total number of transactions under the conditions identified earlier.
The Cisco WAAS device reduces the amount of unnecessary data that traverses the WAN by locally caching data and using compression algorithms on the data the must traverse the WAN. As seen in Figure 18, the Cisco WAAS-enabled network becomes more efficient as less data must traverse the WAN.
With this efficiency, the end-user transaction times are faster and more transactions can occur as shown in the figures below. Figure 19 and Figure 20 indicate the number of transactions that were observed for the same 30 minute cycle.
Because of the size of this WAN link, without Cisco WAAS it became very congested with 40 users performing site navigation and hence some of the transactions failed. This is seen in the both the transaction times chart and in the transaction summary chart. In the transaction times chart, some of the "without Cisco WAAS" average response times are not shown. This is due to some of the transactions (as seen in the transaction summary chart) failing, since the WAN was congested, thus producing no response times. The end-user would receive an application timeout error message.
The transaction times chart also indicates that with Cisco WAAS enabled the WAN link becomes more efficient, allowing transactions to respond to end-user requests. The Cisco WAAS devices remove the congestion by reducing the amount of unnecessary data traffic that traverses the WAN by locally caching data and using compression algorithms on the data the must traverse the WAN. The transaction summary chart also indicates that more transactions are successful and more transactions are allowed.
Figure 19 Transaction Time for 512 Kbps WAN Link
Figure 20 Number of Transactions for 512 Kbps WAN Link—With and Without Cisco WAAS
This section focuses on the network management system (NMS) that was used to monitor and provide results indicating the benefits of the Cisco WAAS optimization. The NMS tool used was NetQoS SuperAgent with NetQoS Collector and Reporter. NetQoS Collector gathers the pre-optimized traffic and reports the data to the NetQoS SuperAgent. The NetQoS SuperAgent provides details on the protocols and applications traversing the network(s), including:
•Data transfer time
•Network round trip time
•Effective network round trip time
•Performance by the server
•Performance by the network
This information provides the baseline of the application under test with valid overall transaction times (end user experience).
NetQoS Reporter gathers the optimized traffic and reports the data to NetQoS Super Agent. NetQoS Super Agent uses the data from the NetQoS Collector (un-optimized) and compares it to the optimized traffic, indicating the benefits of optimization using the Cisco WAAS as shown in the generic samples in Figure 21, Figure 22, and Figure 23.
Figure 21 Benefit of Optimization Using the Cisco WAAS—Application Response Time
Figure 22 Benefit of Optimization Using the Cisco WAAS—Application Data Rate
Figure 23 Benefit of Optimization Using the Cisco WAAS—Link Utilization
NetQoS devices passively listen in by using the rspan feature of Cisco routers and switches. They do not poll servers and hence do not add to the server load. For more information about this tool, see:
Appendix A—Cisco ACE Configuration
Cisco ACE Admin Contextlogin timeout 60line vtysession-limit 100hostname ACE1-Slot3boot system image:c6ace-t1k9-mz.3.0.0_A1_5a.binresource-class Goldlimit-resource all minimum 0.00 maximum unlimitedlimit-resource conc-connections minimum 10.00 maximum unlimitedlimit-resource sticky minimum 10.00 maximum unlimited! Define an access control list to include all IP and ICMP trafficaccess-list anyone line 10 extended permit ip any anyaccess-list anyone line 20 extended permit icmp any anyclass-map type management match-any REMOTE-ACCESSdescription remote access traffic match rule10 match protocol telnet any20 match protocol ssh any30 match protocol icmp any! Define a policy to allow management trafficpolicy-map type management first-match REMOTE-MGTclass REMOTE-ACCESSpermitinterface vlan 82description To Management Networkip address 172.28.196.150 255.255.255.0access-group input anyoneservice-policy input REMOTE-MGTno shutdownft interface vlan 500description FT connectionip address 192.168.50.1 255.255.255.252peer ip address 192.168.50.2 255.255.255.252no shutdownft peer 1heart interval 300heart count 10ft-interface vlan 500ft group 1peer 1no preemptpriority 200associate-context Admininserviceip route 0.0.0.0 0.0.0.0 172.28.196.1! Create a context for SharePointcontext sharepointdescription SharePoint Environmentallocate-interface vlan 82allocate-interface vlan 210-211member Goldsnmp-server community xxxxxx group Network-Monitor! For each fault tolerance group, associate a context with itft group 2peer 1no preemptpriority 200associate-context sharepointinservice
Cisco ACE SharePoint Contextlogin timeout 60!! Define SSL CSR parameterscrypto csr-params MOSSSSLPARMScountry USstate Californialocality SJorganization-name CMOorganization-unit ANSteamcommon-name teamserial-number cisco123!!! Define an access control list to include all IP and ICMP trafficaccess-list ANYONE line 10 extended permit ip any anyaccess-list ANYONE line 20 extended permit icmp any any! Configure different types of probes to check if servers are healthyprobe icmp PINGinterval 2faildetect 2probe tcp IISport 80interval 2faildetect 2passdetect interval 10passdetect count 2! Define parameters for TCP reuse and SSL terminationparameter-map type http TCP-REUSEserver-conn reusecase-insensitive!parameter-map type ssl SSLPARMScipher RSA_WITH_RC4_128_MD5version SSL3! Configure SharePoint servers as real servers and place them in servicerserver host SP1ip address 10.1.211.10inservice!rserver host SP2ip address 10.1.211.11inservice! Define SSL Proxy servicessl-proxy service SSLPROXYkey moss.keycert mosscert1.pemssl advanced-options SSLPARMS! Define server farms for Layer 4 and Layer 7 load balancingserverfarm host MOSSpredictor leastconnsprobe PINGprobe IISrserver SP1inservicerserver SP2inservice!serverfarm host MOSS-FARM2probe PINGprobe IISrserver SP1 80inservicerserver SP2 80inservice! Define various classes for http traffic,! for management traffic,! for traffic destined to VIP with port 80class-map type management match-any REMOTE-MANAGEMENT2 match protocol telnet any3 match protocol icmp any4 match protocol ssh any5 match protocol snmp any6 match protocol http any7 match protocol https any!class-map match-all VIP-HTTP-102 match virtual-address 10.1.210.10 tcp eq www!class-map match-all VIP-SSL-102 match virtual-address 10.1.210.10 tcp eq https! Define various policies to allow management traffic,! load balance the http traffic using session persistence.! to use a previously defined policy and enable the VIP to reply to ICMP requestspolicy-map type management first-match REMOTE-MANAGEMENTclass REMOTE-MANAGEMENTpermit!policy-map type loadbalance first-match VIP-POLICY-10class class-defaultserverfarm MOSS!policy-map type loadbalance first-match VIP-SSL-10class class-defaultserverfarm MOSS-FARM2!policy-map multi-match LB-VIPclass VIP-HTTP-10loadbalance vip inserviceloadbalance policy VIP-POLICY-10loadbalance vip icmp-reply!class VIP-SSL-10loadbalance vip inserviceloadbalance policy VIP-SSL-10loadbalance vip icmp-replyappl-parameter http advanced-options L7-MAPssl-proxy server SSLPROXY! For the client side interface, define IP address for this Cisco ACE, the other Cisco ACE and the shared alias address of the two Cisco ACEs. Load balance traffic destined to VIPinterface vlan 210description Client side vlanip address 10.1.210.5 255.255.255.0alias 10.1.210.4 255.255.255.0peer ip address 10.1.210.6 255.255.255.0access-group input ANYONEservice-policy input LB-VIPservice-policy input REMOTE-MANAGEMENTno shutdown! For the server side interface, define IP address for this Cisco ACE, the other Cisco ACE and the shared alias address of the two Cisco ACEs, load balance.interface vlan 211description Server side vlanip address 10.1.211.2 255.255.255.0alias 10.1.211.1 255.255.255.0peer ip address 10.1.211.3 255.255.255.0access-group input ANYONEservice-policy input REMOTE-MANAGEMENTno shutdown! Default route to the HSRP address on aggregation routersip route 0.0.0.0 0.0.0.0 10.1.220.1
Appendix B—Cisco WAE Configurations
Branch Cisco WAE Configuration! WAAS version 4.0.13 (build b12 Aug 9 2007)! Configure this device to function as a Cisco WAAS Enginedevice mode application-accelerator!!hostname ANS-EDGE!!clock timezone US/Pacific -7 0!!ip domain-name cisco.com!!!primary-interface GigabitEthernet 1/0!!! Connect to the branch routerinterface GigabitEthernet 1/0ip address 10.1.12.2 255.255.255.0exit!! This is the address of interface vlan301 on the branch router.ip default-gateway 10.1.12.1!no auto-register enable!! ip path-mtu-discovery is disabled in Cisco WAAS by default!ip name-server 22.214.171.124!!! Designate the server for network time protocolntp server 10.1.20.1!!wccp router-list 1 10.1.12.1wccp tcp-promiscuous router-list-num 1 password ****wccp version 2!!!snmp-server community ANSwerLab!!!windows-domain netbios-name "ANS-EDGE"!authentication login local enable primaryauthentication configuration local enable primary!!! Enable central managercentral-manager address 10.1.21.2cms enable!!!flow monitor tcpstat-v1 host 10.1.70.11flow monitor tcpstat-v1 enable!tfo tcp optimized-send-buffer 512tfo tcp optimized-receive-buffer 512!!no adapter epm enable!! The application traffic is traversing the WAN using port 80. The default policy configured on the Cisco WAE will be applied. Note that the application configuration can be modified to any port.policy-engine application...classifier HTTPmatch dst port eq 80match dst port eq 8080match dst port eq 8000match dst port eq 8001match dst port eq 3128exitclassifier HTTPSmatch dst port eq 443exit...classifier NetQoSmatch dst port eq 7878exit! Full optimization is applied to the application WAN trafficmap basicname NetQoS classifier NetQoS action optimize full...name Web classifier HTTP action optimize fullname Web classifier HTTPS action optimize DRE no compression none...! End of WAAS configuration
Data Center Cisco WAE Configuration! WAAS version 4.0.13 (build b12 Aug 9 2007)! Configure this device to function as a Cisco WAAS Enginedevice mode application-accelerator!!hostname ANS-CoreWAE!!clock timezone US/Pacific -7 0!!ip domain-name cisco.com!!!primary-interface GigabitEthernet 1/0!!! Connect to the data center WAN edge routerinterface GigabitEthernet 1/0ip address 10.1.20.2 255.255.255.0exit!!! This is the address of interface GigabitEthernet2/0 on data center WAN edge router.ip default-gateway 10.1.20.1!no auto-register enable!! ip path-mtu-discovery is disabled in Cisco WAAS by default!ip name-server 126.96.36.199!!! Designate the server for network time protocolntp server 10.1.20.1!!wccp router-list 1 10.1.20.1wccp tcp-promiscuous router-list-num 1 password ****wccp version 2!!!snmp-server community ANSwerLab!!!windows-domain netbios-name "ANS-COREWAE"!authentication login local enable primaryauthentication configuration local enable primary!!! Enable central managercentral-manager address 10.1.21.2cms enable!!!flow monitor tcpstat-v1 host 10.1.70.11flow monitor tcpstat-v1 enable!tfo tcp optimized-send-buffer 2048tfo tcp optimized-receive-buffer 2048!!! The application traffic is traversing the WAN using port 80. The default policy configured on the Cisco WAE will be applied. Note that the application configuration can be modified to any port.policy-engine application...classifier HTTPmatch dst port eq 80match dst port eq 8080match dst port eq 8000match dst port eq 8001match dst port eq 3128exitclassifier HTTPSmatch dst port eq 443exit...classifier NetQoSmatch dst port eq 7878exit! Full optimization is applied to the application WAN trafficmap basicname NetQoS classifier NetQoS action optimize full...name Web classifier HTTP action optimize fullname Web classifier HTTPS action optimize DRE no compression none...! End of WAAS configuration
Appendix C—Cisco ACE Appliance Configuration and Test Results
Figure 24 Cisco ACE Appliance Configuration
The enterprise branch design shows the Cisco WAE Appliance connected to the local branch router, typically a Cisco Integrated Services Router (ISR). The design provides scalability and availability as compared to installing a Cisco WAAS Network Module within a Cisco ISR as the Cisco ISR must share its resources.
HP Mercury LoadRunner, running on a personal computer in the branch, simulates users that would perform certain routine tasks in the application.
The traffic is redirected to the Cisco WAE through the Web Cache Communications Protocol (WCCP) from the branch router. The Cisco WAE performs the following functions:
•Locally cached—If the data that is being requested is locally cached, the Cisco WAE responds to the requestor with the cached data and requests only required data from the server farm. This allows the WAN to become more efficient as only required data is requested.
•New data—If the data that is being forwarded to the server farm or coming from the server farm, the Cisco WAE performs compression algorithms on the data, allowing for the WAN to become more efficient.
The WAN simulator provide simulations of standard T1. The following simulation was used:
•WAN Type 1 (Intracontinental)
–Bandwidth-1.544 Mbps, ESF, B8ZS
–Loss-Drop one packet in every 1000 packets
For this design, the Cisco ACE Appliance is targeted for a small-to-medium data center (DC). The DC follows the design guidelines found in the Data Center Infrastructure Design Guide found at: http://www.cisco.com/go/designzone.
The design consists of a DC WAN router, a collapsed core/aggregation, access, and the server farm (where the application resides). In this appendix, the focus is on the DC WAN router, aggregation, and the server farm. The core provides routing to and from the DC WAN router and the aggregation. The access layer provides Layer 2 connectivity for the server farms to the aggregation. For larger deployments, one should consider a separate core and aggregation layer, or a one-arm deployment where the Cisco ACE Appliance connects to a Cisco 6500. For more information, refer to:
The DC WAN router performs the same function as the branch WAN router by redirecting traffic to the DC WAE. The DC WAE performs the following:
•Locally cached—If the data that is being requested is locally cache, the WAE responds to the requestor with the cached data and requests only required data from the branch. This allows the WAN to become more efficient as only required data is requested.
•New data—If the data is being forwarded to the branch or coming from the branch, the WAE performs compression algorithms on the data, allowing for the WAN to become more efficient.
Within a Cisco WAAS topology, each Cisco WAE runs a process called central management system (CMS). The CMS process provides SSL-encrypted bidirectional configuration synchronization of the Cisco WAAS Central Manager and the Cisco WAE devices. The CMS process is also used to exchange reporting information and statistics at a configurable interval. When the administrator applies configuration or policy changes to a Cisco WAE device or a group of Cisco WAE devices (a device group), the Cisco WAAS Central Manager automatically propagates the changes to each of the managed Cisco WAE devices. Cisco WAE devices that are not available to receive the update receive the update the next time they become available.
Note Cisco WAEs out of box have a standard set of application variables and ports that are defined for optimization. In this solution no changes need to be made to the standard default configuration of the Cisco WAEs.
The aggregation segment contains the Cisco ACE Appliance. The Cisco ACE Appliance provides the following:
•Virtualization—Device partitioning, where the Cisco ACE has multiple contexts. Each context can be configured for different applications and each context is independent of the other. The Cisco ACE is configured with Admin context and the SharePoint context. Note that the Cisco ACE can support up to 20 contexts (dependent on the license).
•Session Persistence—The ability to forward client requests to the same server for the duration of the session. The SharePoint application requires cookie stickiness session persistence.
•Transparent Interception—Performs a NAT function to conceal the real server IP address that is residing in the server farm. The SharePoint context is configured with a Virtual IP (VIP) that provides a single address for the users to connect to the server farm with. This allows users to access the SharePoint application by placing a single IP address in the Web browser.
•Health monitoring—Used to track the state of the server and determine its ability to process connections in the server farm. The SharePoint context used TCP probes to verify if the SharePoint Web servers were available to process application connections.
The Cisco ACE Appliance provides load balancing of the traffic bound to the server farm using one of the following methods:
•Weighted Round Robin
Round Robin was used to provide load balancing for the SharePoint Web Servers. Round Robin sequentially alternates the connections among available servers.
ACE 4710 Configuration
Admin Contextinterface gigabitEthernet 1/1 <--Connection to Client-side networkdescription CLIENT-WANswitchport access vlan 70no shutdowninterface gigabitEthernet 1/2shutdowninterface gigabitEthernet 1/3 <--Connection to Server-side networkdescription DC-LANswitchport trunk allowed vlan 71no shutdowninterface gigabitEthernet 1/4shutdownresource-class STICKYlimit-resource all minimum 0.00 maximum unlimitedlimit-resource sticky minimum 10.00 maximum unlimitedboot system image:c4710ace-mz.A1_7.0.5.binlogin timeout 0class-map type management match-any MGMT-TRAFFICdescription "allowed mgmt traffic to ACE"2 match protocol http any3 match protocol https any4 match protocol icmp any5 match protocol ssh any6 match protocol telnet any7 match protocol xml-https anypolicy-map type management first-match REMOTE-MGMTclass MGMT-TRAFFICpermitinterface vlan 70 <--Management Interfaceip address 188.8.131.52 255.255.255.0service-policy input REMOTE-MGMTno shutdownip route 0.0.0.0 0.0.0.0 184.108.40.206context Sharepoint <--Creation of Context for Sharepointallocate-interface vlan 70-71member STICKYusername admin password 5 $1$faXJEFBj$TJR1Nx7sLPTi5BZ97v08c/ role Admin domain default-domainusername www password 5 $1$faXJEFBj$TJR1Nx7sLPTi5BZ97v08c/ role Admin domain default-domain
SharePoint Contextlogging console 7logging monitor 7login timeout 0access-list NON-RESTRICTIVE line 8 extended permit ip any anyaccess-list STRICT line 8 extended permit tcp any eq www host 220.127.116.11access-list STRICT line 16 extended permit tcp any eq https host 18.104.22.168probe icmp L3 <--Layer 3 Health Probeinterval 10passdetect interval 10probe http SP-HTTP <--Layer 7 Health Probedescription "Test HTTP request to Portal Page"interval 20passdetect interval 10request method get url /default.aspx <--URL checked on Serversexpect status 200 200rserver host SP-2 <--Assignment of SharePoint Web Serversdescription "SHAREPOINT server 2"ip address 22.214.171.124probe L3inservicerserver host SP-3description "SHAREPOINT server 3"ip address 126.96.36.199probe L3inserviceserverfarm host SHAREPOINT <--Assignment of SharePoint Serverfarmrserver SP-2 80probe SP-HTTPinservicerserver SP-3 80probe SP-HTTPinservicesticky http-cookie ACE_SHAREPOINT ACE-COOKIE <--Cookie Insertcookie insertserverfarm SHAREPOINTclass-map type management match-any REMOTE-ACCESSdescription "Define Allowed Mgmt Traffic"2 match protocol http any3 match protocol https any4 match protocol telnet any5 match protocol ssh any6 match protocol xml-https any7 match protocol snmp any8 match protocol icmp anyclass-map match-any SP-VIP-HTTP <--VIP for Port 80 Traffic to Servers2 match virtual-address 188.8.131.52 tcp eq wwwpolicy-map type management first-match REMOTE-MGMTclass REMOTE-ACCESSpermitpolicy-map type loadbalance first-match SHAREPOINTclass class-defaultsticky-serverfarm ACE-COOKIE <--Load balance with Cookie Stickinesspolicy-map multi-match SHAREPOINT-LBclass SP-VIP-HTTPloadbalance vip inserviceloadbalance policy SHAREPOINTloadbalance vip icmp-reply activeinterface vlan 70description CLIENT-WANip address 184.108.40.206 255.255.255.0access-group input NON-RESTRICTIVEservice-policy input REMOTE-MGMTservice-policy input SHAREPOINT-LBno shutdowninterface vlan 71description DC-LANip address 220.127.116.11 255.255.255.0access-group input NON-RESTRICTIVEno shutdownip route 0.0.0.0 0.0.0.0 18.104.22.168
Microsoft SharePoint server was tested for site navigation, document upload, and document download. Using the HP Mercury LoadRunner testing tool, typical scenarios were designed in which multiple users would perform each type of transaction. Cisco's application delivery solution comprised the hardware load balancer ACE 4710 appliance and Cisco's WAN optimization device, WAAS, delivered up to 99% performance improvement in the transaction response times.
While the ACE load balancer optimally diverted the Web traffic to multiple SharePoint web servers, WAAS reduced traffic over the WAN link (refer to WAAS documents for details).
In the site navigation test, 40 users accessed the SharePoint application simultaneously. They navigated through various parts of the application. As shown in Figure 25, there was a 68% to 94% improvement in the transaction response times.
Figure 25 Site Navigation Test Results
In this test, 40 users uploaded 512kb, 1mb, and 5mb files to the SharePoint servers. The ACE appliance balanced this traffic across the two SharePoint application servers. As shown in Figure 26, WAAS demonstrated up to 99% savings in response time.
Figure 26 File Upload Test Results
When 40 users tried to download various types of files, the ACE redirected them to two application servers while WAAS optimized the file download by up to 99%.
Figure 27 File Download Test Results
Enterprise Data Center Wide Area Application Services (WAAS) Design Guide:
SharePoint planning documents:
•Planning and architecture for Office SharePoint Server 2007:
Some relevant sections:
–Design logical architecture:
–Plan for performance and capacity (Office SharePoint Server)
Cisco Advanced Services
Cisco Services Help Accelerate and Optimize ANS Deployments
Application deployments are complex projects. Cisco Services can help mitigate the risk of making changes to the environment and accelerate deployment of Cisco ANS solutions. Our product and technology expertise is constantly enhanced by hands-on experience with real-life networks and broad exposure to the latest technology and implementations. Cisco uses leading practices to help our customers define their IT and business requirements and help them deliver fast, secure and highly available application access in a scalable environment.
•The Cisco Application Control Engine Planning and Design Service helps customers accelerate deployment of a Cisco ACE solution for fast, secure application access in a scalable environment.
•The Cisco Application Control Engine Optimization Services help customers continuously update and optimize their Cisco Application Control Engine solution as their applications delivery environment changes.
•The Cisco Wide Area Application Services Planning and Design Service helps customers accelerate deployment of Cisco WAAS solutions and improve application responsiveness across their wide area networks.
•The Cisco Wide Area Application Services Optimization Services help customers maintain or improve application responsiveness across wide area network as their business changes and grows.
•The Cisco Application Profiling Service helps customers host and manage applications more effectively while preserving application performance, security, and availability.
Cisco ANS Services:
Cisco Validated Design
The Cisco Validated Design Program consists of systems and solutions designed, tested, and documented to facilitate faster, more reliable, and more predictable customer deployments. For more information visit www.cisco.com/go/validateddesigns.
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.
CCDE, CCENT, Cisco Eos, Cisco Lumin, Cisco Nexus, Cisco StadiumVision, Cisco TelePresence, the Cisco logo, DCE, and Welcome to the Human Network are trademarks; Changing the Way We Work, Live, Play, and Learn and Cisco Store are service marks; and Access Registrar, Aironet, AsyncOS, Bringing the Meeting To You, Catalyst, CCDA, CCDP, CCIE, CCIP, CCNA, CCNP, CCSP, CCVP, Cisco, the Cisco Certified Internetwork Expert logo, Cisco IOS, Cisco Press, Cisco Systems, Cisco Systems Capital, the Cisco Systems logo, Cisco Unity, Collaboration Without Limitation, EtherFast, EtherSwitch, Event Center, Fast Step, Follow Me Browsing, FormShare, GigaDrive, HomeLink, Internet Quotient, IOS, iPhone, iQ Expertise, the iQ logo, iQ Net Readiness Scorecard, iQuick Study, IronPort, the IronPort logo, LightStream, Linksys, MediaTone, MeetingPlace, MeetingPlace Chime Sound, MGX, Networkers, Networking Academy, Network Registrar, PCNow, PIX, PowerPanels, ProConnect, ScriptShare, SenderBase, SMARTnet, Spectrum Expert, StackWise, The Fastest Way to Increase Your Internet Quotient, TransPath, WebEx, and the WebEx logo are registered trademarks of Cisco Systems, Inc. and/or its affiliates in the United States and certain other countries.
All other trademarks mentioned in this document or Website 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. (0807R)