Downloads |
Table Of Contents
Enterprise Branch Security Design Guide
Threat Defense Detection and Mitigation
Configuration and Implementation
Threat Defense Detection and Mitigation
Appendix A—Cisco Platforms Evaluated
Appendix B—Cisco IOS Releases Evaluated
Access Router #1 Configuration
Access Router #2 Configuration
Access Router #1 Configuration
Access Router #2 Configuration
Stackwise Switch Master Configuration
Appendix D—References and Recommended Reading
Enterprise Branch Security Design Guide
This design chapter offers guidelines and best practices for securing the enterprise branch. The following three branch profiles are described to address various customer requirements balancing cost, security, availability, and manageability:
•
Single-tier
•
•
In each profile, the concepts of high availability, infrastructure protection, secure connectivity, and threat defense are addressed. This chapter lays the foundation for integration of advanced services into the enterprise branch architecture.
Introduction
This design chapter evaluates securing an enterprise branch as it pertains to the Enterprise Branch Architecture framework. The Enterprise Branch Architecture is one component in the overall Cisco Service Oriented Network Architecture (SONA) that provides guidelines to accelerate applications, business processes, and profitability. Based on the Cisco SONA framework, the Enterprise Branch Architecture incorporates networked infrastructure services, integrated services, and application networking services across typical branch networks, as shown in Figure 1.
Figure 1 Enterprise Branch Architecture Framework
This design chapter focuses on building single-tier, dual-tier, and multi-tier branch profiles. Each profile provides guidelines for LAN and WAN deployment, network fundamentals such as routing and high availability, and guidance on how to secure a branch through infrastructure protection, secure connectivity, and threat defense. The three profiles establish a foundation to provide guidance as various integrated services are added to the Enterprise Branch Architecture.
This design chapter begins with an overview, which is followed by design recommendations. In addition, configuration examples are also presented. Each service is described in detail and then shown in the three profiles to provide complete guidance on how to secure a branch with the intention of adding various advanced services in the future.
Design Overview
The topology of a typical branch network varies greatly between one enterprise customer and another. Each branch network design reflects the size, industry specific, location, and cost constraints of the customer. Regardless of network architecture, there is a set of common branch networking elements that include routers, switches, and, optionally, dedicated security appliances to provide network connectivity. Users at each branch contain a combination of phones, laptops, and video equipment to run various applications. Point-of-sale terminals, badge readers, and video devices may also require network access. Access points and call processing equipment might be required in branches that require mobility and centralized voice in their network.
Designing a branch network may not appear to be as interesting or exciting as designing an IP telephony network, an IP video network, or even designing a wireless network. However, emerging applications such as these are built on a branch foundation. The Enterprise Branch Architecture introduces the concept of three branch profiles that incorporate the common branch network components. These three profiles are not intended to be the only architectures recommended for branch networks, but rather a representation of various aspects branch networks need to include. These profiles are used as the baseline foundation in which all the integrated services building blocks and application networking services are built. This design chapter builds the foundation through the three profiles.
This design chapter provides an overview of the three profiles tested. The profile approach is meant to provide guidance for using several network architectures to allow the reader to mix and match between profiles without having to test every single branch architecture available. The following fundamental services are provided in this chapter:
•
•
•
•
As each service is defined in detail, the implementation of each service in each profile is discussed. In the end, the three profiles provide guidance on how to secure a branch with high availability using the common branch networking components.
Design Components
The design components for this design chapter comprise the networked infrastructure layer of the overall Enterprise Branch Architecture Framework. From the common network elements, three profiles are presented. The three profiles tested are the single-tier, dual-tier, and multi-tier branch profiles, as shown in Figure 2. Each profile is discussed in greater detail in the following sections.
Figure 2 Three Profiles—Overview
Single-Tier Branch Profile
The single-tier branch profile consists of a fully integrated, one-box solution. All network functions such as LAN or WAN that are necessary for a branch exist in a single tier or device, as shown in Figure 3.
Figure 3 Single-Tier Branch Profile
Typically, the access router consists of an Integrated Services Router (ISR) with an integrated switch module installed in one of the network module slots. The WIC slots provide WAN connectivity to either a campus, headquarters, or the Internet. In Cisco testing, the single-tier branch profile used a T1 link to the Internet, with ADSL through the Internet as a backup link. This profile was chosen to demonstrate a one-box, all-inclusive branch office solution. The benefit of the single-tier branch profile is a single device solution. The drawbacks to this profile include no box redundancy for high availability, and the limited number of users because of the limited number of LAN ports per network module. This profile takes advantage of various Cisco IOS features. However, the probability of reaching the maximum router CPU is greater in this profile. Although during this phase of enterprise branch testing, the CPU utilization remained below 85 percent for the ISR portfolio, it is expected that as more services are added in the future, some ISR platforms may run out of CPU. This profile is intended for smaller enterprise branches that wish to integrate as many advanced services as possible into a single management platform solution.
Dual-Tier Branch Profile
The dual-tier branch profile provides a two-layer architecture consisting of two access routers connected to an external Catalyst switch, as shown in Figure 4.
Figure 4 Dual-Tier Branch Profile
The access routers tested were the ISR portfolio, and the Catalyst 3750 switch was used. Although the Catalyst switch is configured as a Layer 2 device similar to the integrated EtherSwitch module in the single-tier branch profile, the device is external to the access router. The access routers use the integrated Gigabit Ethernet ports to attach to the switch and the WIC slots for WAN connectivity. Connectivity to the campus or headquarters is provided through a Frame Relay link.
It is also assumed in this profile, as it is in the single-tier branch profile, that all services reside in the headquarters. This profile does add an additional layer of devices. By having dual access routers, each with a WAN connection to the headquarters and a LAN connection to the external desktop switch, this branch architecture is more redundant and provides more high availability than the single-tier branch profile. The dual-tier profile is designed to resemble a significant portion of the current branch architectures available in the enterprise market. Separated LAN functionality from the access router providing WAN connectivity is common. This profile intends to show a migration path for legacy branches to integrate advanced services residing in their current branch architecture without having to forklift their current equipment.
Multi-Tier Branch Profile
The multi-tier branch profile separates network functionality into a separate device layer. The tiers in this profile are WAN termination, firewall functionality, services termination, and LAN functionality, as shown in Figure 5.
Figure 5 Multi-Tier Branch Profile
Compared to the other two profiles, each network function in the multi-tier branch profile exists on a dedicated device. WAN termination is provided through two access routers that are connected to a pair of Adaptive Security Appliance (ASA) security appliances that provide firewall functionality. The ASAs are connected to a second set of access routers that will host advanced services such as mobility and IPC communications in future phases. A desktop switch is connected to these access routers for LAN termination. In addition to having a separate device layer for services, each device has redundancy for failover scenarios. Although the separated functionality and dual device redundancy results in the most complexity and expense of the three profiles presented, the benefits are redundancy, availability, and router and switch CPU utilization. Because network services are implemented on distinct devices, each with dedicated CPU resources, the likelihood of exhausting the CPU is less than the other profiles. Also, LAN users can be easily added because the desktop switches are configured in a Stackwise topology. This profile is intended for large enterprise branch architectures and small campus environments. For this design guide, all services reside across the WAN at the headquarters. As more and more services are added to the Enterprise Branch Architecture testing, this profile is ideal for hosting the services at the branch that require high availability and resiliency.
Design Component Summary
Three profiles established in the Enterprise Branch Architecture have varying ranges of cost, management, and resiliency. The single-tier branch profile provides a fully integrated solution that is cost-effective and easy to manage at the expense of high availability and redundancy. The dual-tier branch profile separates LAN and WAN functionality and provides greater availability and redundancy. However, there are additional costs to consider and more devices to manage overall. The multi-tier branch profile provides the least integrated functionality solution with the most devices to manage. However, this solution provides the most availability, redundancy, and resiliency of any of the other profiles. The testing results of all three profiles are included in this design chapter to provide a template for a specific customer branch architecture. It is fully expected that many branch architectures will contain some parts of each profile presented. This design chapter is organized to address each network service individually. Under each section for a specific network service, all three profiles are presented, and guidance for each of the three scenarios is provided. The profile approach for each individual network service offers the most flexibility and modularity to provide the most guidance for integrating advanced services into most types of required branch architecture.
Design and Implementation
This section addresses each of the three profiles described in Design Components, using several of the integrated services building blocks as described in the overall Enterprise Branch Architecture. This section discusses the following services:
•
•
•
•
Each service building block is described as it applies to each profile, and specific implementation issues for each service and profile are discussed.
WAN Services
WAN services provide the foundation for the Enterprise Branch Architecture to connect to the campus or data center core using an Internet service provider (ISP), a traditional service provider network, private leased lines, or some combination of these technologies. The branch may also obtain Internet access directly (split tunnel) without first accessing the campus core. The WAN services building block consists of the following three fundamental deployment options, each with its own set of associated attributes, as shown in Figure 6:
•
•
•
The set of attributes associated with each profile influences the use of specific features, and requires specific considerations when designing a branch office. Each of the three profiles address a separate WAN deployment model.
Figure 6 WAN Deployment Models
Internet Deployment Model
An Internet deployment model provides limited separation or segmentation of the enterprise network traffic and, as such, most deployments use IP security (IPsec) for data secrecy, authentication, and integrity. With this deployment model, all traffic traverses an ISP cloud. In most cases, WAN links from the branch terminate on an ISP WAN router and traverse the ISP backbone to the enterprise campus. This technology is very cost-effective because the branch-to-core connection is not sensitive to distance. The enterprise branch connects to the nearest ISP hub through a leased-line or a broadband connection and is then aggregated with other subscriber traffic on the ISP backbone. Subscribers are charged on a fixed rate and are still responsible for administering and maintaining the network equipment and services. However, because the traffic is traversing the Internet, QoS or bandwidth may not be as guaranteed compared to the other deployment models.
The routing control is determined by the ISP and, as such, only IP protocol is supported through the cloud. If non-IP protocol is required from a branch architecture, a tunnelling mechanism such as Generic Routing Encapsulation (GRE) is required. The Internet deployment model is ideally suited for use as the integrated WAN transport on the single-tier branch profile because it is the most cost-effective WAN offering for most customers. The single-tier branch profile uses a T1 link to the campus through an Internet cloud with an ADSL link as a backup link. The ADSL link provides additional costs but also provides some form of failover recovery. The ADSL link can be left out of this profile if cost is more important than increased availability. Traffic from the single-tier branch profile is encrypted, and non-IP traffic is tunnelled to the enterprise WAN edge. The mechanism to secure traffic is addressed in Secure Connectivity. An advantage to the Internet deployment model is that future branch architectures can communicate in an any-to-any inter-site connection, full-mesh topology. However, when considering adding latency and jitter-sensitive services such as voice or video, additional consideration must be taken because the Internet cloud can guarantee latency and QoS, in some instances such as those found in V3PN networks, but at perhaps additional costs, and only from select service providers.
Private WAN Deployment Model
The private WAN deployment model is the traditional hub-and-spoke model that has been deployed in enterprise networks for decades. Traditional Frame Relay or ATM networks are categorized in this deployment model. Data privacy is provided through traffic separation such as Frame Relay data-link connection identifiers (DLCIs) or ATM virtual circuits (VCs). Routing is controlled by the enterprise core network, and both IP and non-IP protocols are supported. No encryption or tunnelling mechanism is required because connectivity is provided at Layer 2, but can be used depending on the exact branch requirements of the customer.
The dual-tier branch profile uses a Frame Relay private WAN deployment model. Each access router has been provisioned to contain a single Frame Relay link to the enterprise WAN edge via a point-to-point T1 link. Separate DLCIs are configured to provide data privacy within the branch and through the external branch cloud. The majority of Frame Relay networks deployed are provisioned by service providers for data transmission services. Frame Relay is implemented in both public carrier-provided networks and in private enterprise networks. In public carrier-provided Frame Relay networks, the Frame Relay switching equipment is located in the central offices of a telecommunications carrier. Subscribers are charged based on their network use but are relieved from administering and maintaining the Frame Relay network equipment and services. In private Frame Relay networks, the administration and maintenance of the network are the responsibilities of the enterprise. All the equipment, including the switching equipment, is owned by the customer. The actual implementation of a Frame Relay network is the same regardless of being public or private; however, the cost and ownership are factors.
MPLS Deployment Model
The MPLS deployment model provides the following beneficial applications: MPLS virtual private network (VPN), traffic engineering, and QoS. MPLS is a packet-forwarding technology that uses labels to make data forwarding decisions. MPLS label forwarding is performed with a label lookup for an incoming label, which is then swapped with the outgoing label and finally sent to the next hop. Labels are imposed on the packets only once at the edge of the MPLS network, and removed at the other end to provide data privacy across the MPLS network. Traffic engineering is enabled through MPLS mechanisms that allow traffic to be directed through a specific path, which may not necessarily be the least expensive path in terms of routing protocol metrics. QoS techniques are implemented to ensure that latency-sensitive traffic types are given priority over less important traffic in transit of the network. QoS gives the network administrator the capability to ensure that VoIP or video latency requirements are met. Only IP traffic traverses an MPLS cloud, so a tunnelling mechanism is required for non-IP traffic. The design intent of the multi-tier branch profile is high availability and resiliency. The MPLS deployment model was chosen based on the benefits the MPLS technology provides compared to the other deployment models. Many enterprise customers are connecting with two MPLS service providers at the branch and head-end campus to isolate themselves from the failure of a single MPLS network.
LAN Services
LAN services provide end device connectivity to the corporate network within the branch office. With the convergence of services onto a single network infrastructure, devices such as computers, telephones, video cameras, and so on, all require the connection to the corporate network over the LAN. Following are the three prominent physical configurations for LAN connectivity:
•
•
•
These configurations are shown in Figure 7.
Figure 7 Prominent Physical Configurations for LAN Connectivity
Each of these configurations have their own set of advantages and disadvantages, and are discussed in the three profiles defined.
The single-tier branch profile uses the access router with an integrated switch configuration. This profile is intended for smaller branch offices that do not require numerous LAN endpoints. This LAN configuration offers all the switching functionality as does any external desktop switch integrated into a one-box solution. The number of users for a branch network deploying this configuration is limited to the number of ports an access router can support. Table 1 shows the maximum switch ports per platform.
In the single-tier branch profile, the integrated switch is configured as a Layer 2 device using the internal backplane connector as the trunk port to the access router. VLANs are configured for data privacy, but only one IP address is required with the switch as a Layer 2-only device. Inline power is supported in this configuration and full Cisco Catalyst features are supported. The advantage to this design is a one-box solution, which means lower total cost of ownership and a single device for management. The disadvantage is the limited number of ports.
The dual-tier branch profile uses the access router connected to a physically separate Cisco Catalyst switch as a Layer 2-only switch. This LAN configuration in terms of feature parity is the same as the configuration used in the single-tier branch profile. The only differences are that the switch is a separate device, and a cable attaching the access router and the switch is required. This cable can be configured as an EtherChannel or a trunk. In the dual-tier branch profile, the connection to the access router is via a trunk port. Spanning tree does not need to be enabled in this profile to avoid loops because there are only two trunk ports to the access router from the external switch. Inline power is provided, depending on the model of the switch chosen. Additional switchports can be added easily, or a larger switch chassis can be used. The disadvantages to this LAN configuration are an additional device to manage and additional costs of purchasing a separate device. This LAN configuration was chosen for the dual-tier branch profile to provide an additional tier of hardware for each network function, and for medium-size branch deployments where more users are required than the fully integrated configuration without the complexity of the Layer 3 services provided by an external Cisco Catalyst switch. However, if Layer 3 services are eventually needed, the equipment is already in place to provide the most flexibility for future growth.
The access router integrated with Cisco Catalyst switches in a Stackwise technology configuration is leveraged in the multi-tier branch profile. Cisco Stackwise technology provides a method of collectively using the capabilities of a stack of switches. The switches are united into a single logical unit via special stack interconnect cables that create a bi-directional closed-loop path. The stack behaves as a single switching unit that is managed by a master switch elected from one of the member switches. The master switch automatically creates and updates all the switching and routing tables. A working stack can accept new members or delete ones without service interruption. Because of the lack of service interruptions provided by the closed loop created in the stack, this LAN configuration is ideally suited for the multi-tier branch profile. The multi-tier branch profile is mainly focused on availability and resiliency, and the Stackwise technology provides this benefit. The Cisco Catalyst switches chosen are configured as Layer 3 devices. Routing decisions are therefore made in the switches. Inline power is provided depending on the exact Cisco Catalyst switch model chosen. The advantages to this design are high availability and resiliency as well as the ability to add more users without service interruption. The disadvantage of this configuration is that the total amount of devices to manage increases as well as the cost of each additional device.
As with the WAN deployment models, the LAN configurations chosen for each profile are not meant to be the only configurations possible. Each profile can interchange any of the LAN configurations. The LAN configurations chosen for this design chapter for each profile is meant for guidance, but can be deployed in any profile depending on the exact customer requirements. For more in-depth LAN deployment options as they refer to generic LAN designs rather than a profile approach, see the following URL: http://www.cisco.com/en/US/netsol/ns742/networking_solutions_program_category_home.html.
For further details, see the following URLs:
•
•
Network Fundamentals
Network fundamentals refer to the basic services that are required for network connectivity. These services include high availability, IP addressing and IP routing, and QoS. Regardless of which WAN or LAN deployment model is chosen for a branch architecture, network fundamentals are required to provide a foundation for any service to be overlaid onto the branch network.
High Availability
High availability is crucial for modern branch architectures. Remaining productive during a network failure is extremely important for all aspects of a network, and especially for branch networks. There are several aspects of high availability, and the three profiles address each one.
The single-tier branch profile demonstrates a dual WAN link to the enterprise WAN edge for availability, as shown in Figure 8.
Figure 8 Single-Tier Branch Profile High Availability
The primary link to the enterprise WAN edge is through the T1 link with an ADSL link for backup. Very similar to legacy networks that used modems for dial backup, the ADSL link is used only when the primary link fails. The single-tier branch profile illustrates the easiest and most cost-effective means for a branch to provide network resiliency. Although this profile integrates all services into a single device for ease of management and lowered total cost of ownership, this profile does not provide any mechanism for device failover. If the access router fails, the only means for regaining productivity is to replace the failed device. This topology is attractive for small branches because a WAN outage is more common than a branch failure. It provides reasonable availability and at less cost than installing dual devices. Although there is no device redundancy in this profile, an inexpensive way to provide resiliency is through a backup ADSL link. Dual WAN link redundancy provides alternative paths to the enterprise WAN edge in case of an ISP or WAN cloud failure.
The dual-tier branch profile builds upon the single-tier branch profile for added availability. In addition to two Frame Relay links to the enterprise WAN edge, there are also dual access routers for device failure, as shown in Figure 9.
Figure 9 Dual-Tier Branch Profile High Availability
The external desktop LAN switch also has a link to each access router. Hot Standby Routing Protocol (HSRP) is used between the access routers for resiliency. One path from the switch and the router is configured as the primary path, with the other path set in standby. If the primary path fails, the secondary path takes over. The primary path can fail through a bad cable, a bad port on the LAN switch or access router, or if the access router connected to the primary path fails. In one of these conditions, the standby router becomes active and network connectivity is resumed. The dual-tier branch profile provides many layers of redundancy. HSRP provides a failover path if one of the access routers fails. Having dual access routers provides a device backup mechanism within this single geographical location, and the dual Frame Relay links provide a failover mechanism in case of an external WAN cloud failure. The only aspect of this profile that is not resilient is the single LAN switch. This topic is addressed in the multi-tier branch profile.
Network uptime is crucial for enterprise networks. However, many branch networks cannot justify the costs associated with a fully redundant and resilient network. The multi-tier branch profile illustrates this type of network. The high availability configuration is shown in Figure 10.
Figure 10 Multi-Tier Branch Profile—High Availability
At every layer in the network, there is an alternate path mechanism for failover. This profile is suited for branches that require availability approaching 100 percent. At the very least, branch networks that do not have the resources to provide as much resiliency and redundancy can take from this profile for the areas of resiliency and redundancy that they do have in their own customer network. To protect against external WAN failures to the enterprise WAN edge, dual WAN links are provided. There are backup devices to recover from single device failures. The access routers use routing and HSRP to recover from a single device failure. The ASA firewalls use the stateful failover firewall feature. The external Catalyst switches are configured in a stack with the integrated EtherSwitch module in the access routers using the Stackwise technology. This technology is devised so that the switches in the stack are in a closed loop. If a Catalyst switch fails, the stack loops to wrap away from the failed device. Every available aspect of high availability is described in this profile. Although this profile can guarantee the most uptime of any of the three profiles, this profile is also the most difficult to design. Extra precautions for addressing and routing must be considered.
Availability must be considered in any network design, which is why high availability is a network fundamental. The three profiles illustrate the various means for resiliency and redundancy in a branch network. Each of these mechanism can be interchanged to meet a specific customer design.
IP Addressing and IP Routing
Cisco offers a broad portfolio of IP routing technologies. The key protocols that are especially suited for branch networks are EIGRP, OSPF, and policy-based routing (PBR). All routing protocols share common attributes and goals of stability, availability, manageability, fast convergence, and high performance.
EIGRP is used in the configuration examples. OSFP can also be substituted if the customer prefers this routing protocol.
The single-tier branch profile uses EIGRP in the access routers for access to the enterprise WAN edge. Default route and floating static routing are used for WAN failover detection to actively change routes from the primary T1 link to the campus to the ADSL link to the campus. PBR is used to provide non-split tunnelling to the campus; optionally, PBR can be used to avoid split tunnelling while allowing DMVPN spoke-to-spoke. More information on split tunneling is discussed in Secure Connectivity. Figure 11 shows a summary of the routing for the single-tier branch profile.
Figure 11 Single-Tier Branch Profile Routing
The dual-tier branch profile uses EIGRP in the access routers for access to the campus or data center. HSRP is used between the access routers for failover. HSRP provides network redundancy for IP networks, ensuring that user traffic immediately and transparently recovers from first hop failures in network edge devices or access circuits. The routing design for the dual-tier branch profile is designed for all data traffic to traverse through the 2821-1 and all voice traffic to traverse through 2821-2 during normal operations. This design splits the path of both voice and data traffic so that no single access router routes both types of traffic except during a failover condition. Figure 12 shows the dual-tier branch profile routing design.
Figure 12 Dual-Tier Branch Profile Routing
The multi-tier branch profile is the most difficult to design because the ASA firewall does not currently support EIGRP as a routing protocol; it supports only OSPF, RIP, and static. Rather than using OSPF, which the ASA does support, efforts have been made to work around the lack of support of EIGRP using "object tracking route" in this design. Rather than turning off routing capabilities on the ASA firewall and making the ASA firewall function as a Layer 2 transparent firewall or redistributing EIGRP into OSPF and vice versa, static routing is used. The access routers connected into the Catalyst stack all use EIGRP for LAN or inside routing decisions. Object tracking is configured on the links of the access routers connecting to the ASA. Therefore, four interfaces are configured with object tracking to overcome the EIGRP routing dead zone the ASA provides because this device does not support EIGRP. The access routers connecting to the MPLS cloud have EIGRP running for traffic routability between the branch network and the campus network. Figure 13 illustrates how routing is applied for the multi-tier branch profile.
Figure 13 Multi-Tier Branch Profile Routing
The Enhanced Object Tracking feature is used as a failover mechanism similar to how HSRP is used in the dual-tier branch profile. HSRP tracks interface line-protocol state only. If the line protocol of the interface goes down, the HSRP priority of the access router is reduced, allowing another HSRP router with a higher priority to become active. Object tracking can track the IP routing state of the interface, the line protocol state of the interface, IP route reachability, threshold weight, and threshold percentage. Boolean expressions are used to make failover routing decisions with very minimal reconvergence time.
For more information on all the IP routing protocols, see the following URL: http://www.cisco.com/en/US/products/ps6599/products_ios_protocol_group_home.html
As with all designs, the IP addressing and routing protocols chosen is entirely dependent on an individual customer network. Although this section is meant to show guidance for the three profiles tested, http://www.cisco.com/en/US/netsol/ns742/networking_solutions_program_category_home.html provides much documentation for this topic as reference material as well.
Quality of Service
Each of the three profiles chosen serve as a transport for delay-sensitive voice, bandwidth-intensive video, and data applications. QoS is a Cisco IOS software feature that helps manage delay, delay variation (jitter), bandwidth, and packet loss on a network to guarantee efficient, predictable services for business-critical applications. There are various ways to enable QoS on a network. All three profiles use these methods in the same way, so there is no need to individually address each profile. QoS can be enabled on access routers and both the external Catalyst switches and the integrated EtherSwitch service module. The major categories of QoS tested in this design chapter are as follows:
•
•
•
•
•
Figure 14 shows a summary of these QoS categories and where they are placed in a typical branch network.
Figure 14 QoS Strategy in a Typical Branch Network
Although IP telephony is not explored in this design chapter, IP phones are supported in all three profiles. The configurations shown in this guide have been tested with Cisco IP phones in a distributed call processing model. The branch router must be configured to provide QoS support for either a distributed or centralized call processing model.
Packet classification allows traffic to be associated with a priority level or class of service. Packets are selected from a variety of methods ranging from simple the input interface, to access control lists (ACLs), to multi-packet classification using Network-Based Application Recognition (NBAR). NBAR classifies the IP traffic by application level protocol by monitoring the control flows of an application to be able to also correctly classify any new resulting flows. Classification is the first component of Modular QoS CLI (MQC) to allow for clear separation of classes, from the policy applied on the classes to the application of a QoS policy on an interface or subinterface on an access router or switch. Each profile uses NBAR and ACLs to classify traffic. Packets were marked using Layer 2-802.1p/Q, Layer 3- IP precedence, and Differentiated Services Code Point (DSCP) using the policy framework component of MQC.
Weighted random early detection (WRED) algorithm provides for congestion avoidance on network interfaces by providing buffer management and allowing TCP traffic to throttle back before buffers are exhausted. This helps avoid tail drops and global synchronization issues, thereby maximizing network utilization and TCP-based application performance.
Queuing techniques such as weighted fair queuing (WFQ), class-based weighted fair queuing (CBWFQ), low latency queuing (LLQ), and modified deficit round robin (MDRR) are necessary to ensure that critical applications get forwarded even during network congestion. Real-time applications such as voice or video that need to be forwarded with the least latency and jitter use LLQ. Non-delay sensitive traffic can use CBWFQ or MDRR.
Traffic entering a network can be conditioned by using a policer or a shaper. A policer enforces a rate limit while a shaper limits the traffic flow to a specified rate using buffers.
QoS can also provide network security by using scavenger class QoS. The scavenger class QoS strategy identifies known worms and attacks. In a branch network, the end user is a device located on the local LAN residing on a Catalyst switch LAN port. Other traffic patterns from that end user that are considered "unusual" or as "normal traffic but at an unusually high rate" may be marked as Scavenger Class-CS1 in the DSCP field and allowed to pass through the switch. Through the use of the scavenger class, QoS can be used as a security mechanism to limit the arrival rate of any traffic that is destined for the firewall or Intrusion Prevention System (IPS) configurations.
Table 2 summarizes the QoS categories tested in this design chapter and the Cisco IOS features used.
This QoS section provides an overview of the key categories shown in the configuration section of this design guide. For more information, see the Enterprise QoS Solution Reference Network Design Guide Version 3.3 at the following URL: http://www.cisco.com/en/US/docs/solutions/Enterprise/WAN_and_MAN/QoS_SRND/QoS-SRND-Book.html.
Security Services
Security services help protect the device and network from intrusion, tampering manipulation (also called data integrity), secure data transport, and denial of service (DoS). The key categories of security services are the following:
•
•
•
Infrastructure Protection
Infrastructure protection provides proactive measures to protect the infrastructure devices; in this case, Cisco IOS Software-based routers, switches, and appliances, from direct attacks as well as indirect attacks.
Infrastructure protection assists in maintaining network transport continuity and availability. Regardless of the profile chosen, the same methods for infrastructure protection apply. Rather than individually addressing each profile in detail, infrastructure protection applies to all the network components in the branch network. That is, the same infrastructure protection methods apply to access routers, switches, and security appliances. To protect these devices, the following methods are used:
•
•
•
•
•
•
For more information on infrastructure protection techniques, see the following URL: http://www.cisco.com/en/US/docs/solutions/Enterprise/WAN_and_MAN/IPSNGWAN.html.
Secure Connectivity
Secure connectivity protects against information threat or alteration of end user data over untrusted transport mediums. The level of network security that is deployed in a branch depends on the WAN type and deployment model chosen, as shown in Figure 15.
Figure 15 Secure Connectivity Options
In a typical enterprise branch, the WAN types are generally cable/DSL for smaller branches, T1/E1 for medium branches, and T3/E3 for larger branches. The typical WAN deployment models for these WAN types are Internet, private WAN, and MPLS deployment models, as discussed in WAN Services. Both the private WAN and MPLS deployment models as used in the dual-tier branch profile and the multi-tier branch profile respectively, provide a level of secure connectivity through the use of traffic separation. This traffic separation is achieved through Frame Relay DLCIs in the dual-tier branch profile, and MPLS VRFs in the multi-tier branch profile. Traffic is separated from each user; however, the data is not encrypted.
The single-tier branch profile uses the Internet deployment model, which requires a layer of encryption to be applied. Frame Relay and MPLS can run encryption as an additional layer of secure connectivity, although not tested in the dual-tier or multi-tier branch profiles in this design chapter. Network traffic is encrypted through the use of the IPsec standard, which provides a method to manage authentication and data protection between multiple crypto peers engaging in a secure data transfer. The following four ways use the IPsec standard to provide secure connectivity across the WAN:
•
•
•
•
For more information on these four secure connectivity designs using IPsec, see the following URL: http://www.cisco.com/en/US/netsol/ns817/networking_solutions_program_home.html.
The single-tier branch profile uses DMVPN as the secure connectivity method. DMVPN tunnels are configured on both the T1 link and the ADSL link to provide a primary and secondary secure path to the campus. Split tunneling is disabled in this profile so that all traffic must traverse to the campus. Split tunneling is commonly used to allow only corporate traffic to traverse the DMVPN tunnel. All other traffic uses the Internet link outside the branch network. However, to completely encrypt and monitor all traffic leaving the branch network, this design chapter does not allow split tunneling. Disabling split tunneling requires configuring PBR for DMVPN spoke-to-spoke traffic. PBR is required to force routes to each individual spoke because by default, with split tunneling turned off, all traffic is destined for the enterprise WAN edge. More information on spoke-to-spoke DMVPN can be found in the DMVPN design guide mentioned above. The factors to consider are additional security with added routing configuration, or easier routing configuration without complete control over traffic exiting the branch. Both choices are viable and can be used, but the single-tier branch profile in this design chapter chose additional security. Figure 16 shows the secure connectivity design for the single-tier branch profile.
Figure 16 Single-Tier Branch Profile Secure Connectivity
Threat Defense Detection and Mitigation
Threat defense detection and mitigation detects, mitigates, and protects devices against violations and unauthorized events. Each of the three profiles are configured for threat defense. Each network component in the profile determines which method is used. For access routers and security appliances, two of these threat defense mechanisms are through firewalls and IPS, as shown in Figure 17.
Figure 17 Threat Defense Mechanisms for Cisco IOS Routers and Security Appliances
Firewalls provide stateful security and application inspection for each protocol entering or leaving a branch network. A stateful inspection firewall uses a combination of access control with application inspection to ensure that only approved responses get through the firewall. Firewalls can be used through an external appliance such as the ASA in the multi-tier branch profile, or in conjunction with the Cisco IOS Firewall feature set can be used for Cisco IOS access routers as in the single-tier and dual-tier branch profiles.
For more information on the Cisco IOS Firewall Feature Set and the ASA firewall appliance, see the Cisco IOS Firewall Feature Set and the Cisco ASA 5500 Series Adaptive Security Appliances at the following URLs:
•
•
Intrusion protection monitors packets and sessions as they flow through the branch, and scans each packet to match any of the IPS signatures. When a device running IPS, an access router with the Cisco IOS IPS feature set or an external ASA with the IPS feature set loaded, or a standalone IPS sensor detects suspicious activity, it can shunt the offending packets before network security can be compromised. When an IPS signature is matched, one or more of the following actions are taken:
•
•
•
•
For more information on the Cisco IPS feature set, see the Cisco IOS Intrusion Prevention System (IPS) at the following URL: http://www.cisco.com/en/US/products/ps6634/index.html.
The single-tier and dual-tier branch profiles both use the Cisco IOS Firewall and Cisco IOS IPS feature sets embedded into the access routers. The physical WAN links have been designated as the outside interfaces, which means that they are referred to as the "unsecure" network. ACLs are created on these physical WAN links to deny all outside initiated traffic and to provide a modification point for Cisco IOS Firewall inspection dynamic entries. Cisco IOS Firewall inspection creates temporary openings in ACLs at firewall interfaces, which are the DMVPN tunnel links in this instance. These openings are created when specified traffic exits a branch internal network through the firewall. The traffic is allowed back through the firewall only if it is part of the same session of the internal network of the branch as the original traffic that was identified by the Cisco IOS Firewall. Cisco IOS Firewall inspection policies are configured on the DMVPN tunnel links as well.
Only ACLs are configured on the internal interfaces. These interfaces are the LAN interfaces from the access router and the integrated EtherSwitch module for the single-tier branch profile and the link between the external switch and the access router for the dual-tier branch profile. These ACLs only permit or deny specific user traffic. The benefit of these ACLs guarantees is that only allowed user networks can enter a branch network. Unwanted networks are denied.
The multi-tier branch profile is different in that the WAN links are terminated at the first set of access routers, and then this traffic is passed to the ASA for firewall functionality. ACLs can be configured on the WAN termination access routers to deny all outside initiated traffic as a threat defense mechanism. The ASA provides hardware-based, robust firewall capabilities compared to the Cisco IOS Firewall inspection functionality. The same principles apply, but the ASA is functioning only for firewall capabilities. The network behind the ASA is considered the inside secure network, and only ACLs are configured to permit or deny specific user traffic.
IPS is configured on all outside and inside interfaces for all three profiles. Traffic, regardless of whether it is a WAN link to the public or an internal LAN link, is inspected. In testing, IPS signatures 1107 (RFC 1918—Addresses Seen), 2000 (ICMP Echo Reply), and 2001 (ICMP Host Unreachable) are disabled. These signatures can trigger false positives in the lab environment. Running the default IPS signatures loaded with each Cisco IOS release should be sufficient, but updates to the signature file can be made as new signatures are added. A complete list of IPS signatures is located at the following URL: http://www.cisco.com/en/US/prod/collateral/iosswrel/ps6537/ps6586/ps6634/prod_white_paper0900aecd8062ac75.html.
Cisco Catalyst switches have additional mechanisms for threat defense that are applied on a per-port basis, that include the following:
•
•
•
•
Figure 18 shows these mechanisms.
Figure 18 Threat Defense Mechanisms for Catalyst Switches
All three profiles contain Catalyst switches. Whether the switch is integrated, external, or in a stack, the same threat defense mechanisms apply:
•
•
•
•
All the Catalyst switch threat defense mechanisms are used in the three profiles defined, because each profile contains a user base connecting to a Catalyst switch.
Configuration and Implementation
This section addresses how each of the three profiles use several of the integrated services building blocks as described in the overall Enterprise Branch Architecture framework. The services discussed in this design chapter are WAN services, LAN services, network fundamentals, and security services. Each profile and the configurations used for each profile are discussed. Any design issues that need to be considered for each service or each profile are addressed as well. The following figures illustrate the network topology used for each profile. These figures should be referenced as each integrated service is described in more detail.
Figure 19 Single-Tier Branch Profile—Network Topology
Figure 20 Dual-Tier Branch Profile—Network Topology
Figure 21 Multi-Tier Branch Profile—Network Topology
WAN Services
WAN services provide the foundation for the Enterprise Branch Architecture to connect to the enterprise WAN edge and resources in the campus and data center locations via an ISP public or private network, and potentially also Internet access. As a general rule, a branch must have a connection to the WAN to provide a communications channel to the campus to reach the resources found in the campus. As discussed in Design and Implementation, there are three prominent WAN deployment models: Internet, private WAN, and MPLS deployment model. Each of the three profiles test one of the deployment models listed. The following sections show how each of the three profiles are configured for WAN connectivity.
Single-Tier Branch Profile
The single-tier branch profile uses the Internet Deployment model where a primary T1 link is connected through the Internet cloud to reach the campus. For backup, an ADSL link is configured through the Internet. A leased line T1 to the Internet often comes with a fixed and known IP address, and ADSL is likely assigned a dynamic address via DHCP by the Internet service provider. Quite often, this dynamic address is from the RFC 1918 space, and branch traffic uses Network Address Translation (NAT) through this address. The single-tier branch profile is designed to provide a one-box, integrated services solution. WAN connectivity is provided using a VWIC2-2MFT-T1/E1 for T1 connectivity and a WIC-1ADSL for ADSL connectivity. The configuration for an integrated Data Service Unit (DSU) is fairly simple. The line coding and framing are required information. Typically, the line coding is B8ZS and the framing is ESF, but always confirm with the service provider. Either a WIC-1DSU-T1 or a VWIC2-2MFT-T1/E1 can be used. The latter is preferred because it implements more features and positions the deployment for an easy migration to future voice services. As such, the integrated DSU is configured as a T1 controller.
The configurations used for WAN services for the single-tier branch profile are shown in the following examples:
•
controller T1 0/0/0framing esfclock source internallinecode b8zscablelength short 133channel-group 0 timeslots 1-24!interface Serial0/0/0:0ip address 192.168.199.2 255.255.255.248!•
interface ATM0/1/0load-interval 30no atm ilmi-keepalivedsl operating-mode auto!interface ATM0/1/0.35 point-to-pointbandwidth 768no snmp trap link-statuspvc dsl 0/35vbr-nrt 768 768tx-ring-limit 3pppoe max-sessions 5pppoe-client dial-pool-number 1!interface Dialer1bandwidth 768ip address negotiatedip access-group WAN-link inencapsulation pppload-interval 30dialer pool 1dialer-group 1ppp authentication chap callinppp chap hostname soho4@cisco.comppp chap password 7 1316181A0458ppp ipcp dns requestppp ipcp wins request!dialer-list 1 protocol ip permitDual-Tier Branch Profile
The dual-tier branch profile highlights the ability of a legacy private WAN deployment model to implement the integrated services that are defined in the integrated services building block layer of the Enterprise Branch Architecture framework. In the future, this profile will also be used to present a migration path to alternative WANs such as MetroEthernet, while maintaining full integrated services.
A Frame Relay interface configuration is shown for reference. As with the single-tier branch profile, the VWIC2-2MFT-T1/E1 is used to provide WAN connectivity. The WAN services configuration for the dual-tier profile is as follows:
•
interface Serial0/0/0no ip addressencapsulation frame-relay!interface Serial0/0/0.17 point-to-pointip address 10.173.199.2 255.255.255.252frame-relay interface-dlci 17!•
interface Serial0/0/0no ip addressencapsulation frame-relay!interface Serial0/0/0.1 point-to-pointip address 10.173.199.6 255.255.255.252frame-relay interface-dlci 17!Multi-Tier Branch Profile
The multi-tier branch profile uses a service provider-managed MPLS cloud. The service provider hands this connection off as an Ethernet connection. The on-board Ethernet ports on the access routers are used to terminate the MPLS. The WAN connection appears as an Ethernet interface to the access router. Special queuing considerations have to be factored into the design because most MPLS service providers restrict the bandwidth of the connection below the port speed of the router (10 Mbps Ethernet, 100 Mbps FastEthernet, 1 Gbps Gigabit Ethernet). This subject is discussed in detail in Quality of Service.
The WAN services configuration for the multi-tier branch profile are as follows:
•
interface FastEthernet0/0no ip addressduplex autospeed auto!interface FastEthernet0/0.201encapsulation dot1Q 201ip address 10.173.249.1 255.255.255.248!•
interface FastEthernet0/0no ip addressduplex autospeed auto!interface FastEthernet0/0.202encapsulation dot1Q 202ip address 10.173.249.10 255.255.255.248!LAN Services
LAN services provide end device connectivity to the corporate network with the branch office. Typically, end devices such as phones, laptops, and printers connect to a branch network through a switch as they do in a campus. Placement of the switch in the branch network is what distinguishes each of the three profiles. Each profile addresses one of the three prominent LAN configurations but can be interchanged with any of the three profiles.
For all three profiles, VLANs are used to confine traffic into a single logical broadcast domain. VLANs help to segregate the traffic from different endpoints. For instance, voice, video, and data can be segregated by putting the devices into different VLANs. Convergence of voice and data into a single infrastructure lowers the overall cost of ownership of a network, and simplifies administration and maintenance through the elimination of separate voice and data infrastructures. Convergence also implies that to provide the reliability and quality for the voice and data applications, the traffic type has to be identified at the edge of the network so that appropriate QoS parameters can be applied to the traffic.
Cisco switches allow both the voice and data services to be connected to a single physical port. The switch can receive traffic on two VLANs. The first VLAN, called the data VLAN, is sent and received untagged. The second VLAN, called the voice VLAN, is sent tagged with a dot1q header and the voice VLAN to which the device belongs. The tagged packet comes from the IP phone. The data device connected to the IP phone receives and transmits only untagged packets and belongs to the data VLAN. All three profiles show the configuration for voice-only ports, data-only ports, and a data device connected to an IP phone port for completeness.
Single-Tier Branch Profile
The single-tier branch profile uses the integrated switch module with an access router. The links between the access router and the EtherSwitch network module are internal to the chassis. To configure the network module, a session has to be established to the network module from the access router. To establish a session, the Gigabit Ethernet link on the access router has to be configured with an IP address for console access to the network.
When the network module is inserted into a slot, slot 1 of the access router in this example, the interface GigabitEthernet 1/0 on the access router to the interface GigabitEthernet 1/0/2 of the network module is created. The network module can be configured by establishing a session from the access router to the network module. To establish the session, an IP address has to be assigned to the GigabitEthernet interface connecting to the network module from the access router and bringing up the interface by entering a no shut command on the interface. The IP address for this interface does not have to be a routable IP address. For this testing, the IP address of 1.1.1.1 255.255.255.0 was chosen. For more details on configuring the EtherSwitch network module, see the following URL: http://www.cisco.com/en/US/products/ps5854/products_qanda_item0900aecd802a9470.shtml
For this profile, ports were randomly selected as data only, voice only, or a data device connected to an IP phone port to show all three configurations. Depending on the exact customer requirements, ports can be chosen as any of the above. The data device connected to an IP phone ports are Interface FastEthernet 1/0/1-1/0/21, and a data-only port is Interface FastEthernet 1/0/22. VLANs were configured for each type of port: VLAN 193 for data and VLAN 200 for phones. VLANs that have no members or devices attached to the VLAN interface remain in an UP-DOWN status until at least one host or interface joins the VLAN. The trunk port between the switch and the access router is the internal Interface GigabitEthernet 1/0/2. Following are the LAN configurations for the single-tier branch profile:
•
ip dhcp pool data_lannetwork 10.173.193.0 255.255.255.128dns-server 10.59.138.4default-router 10.173.193.1!ip dhcp pool voice_lannetwork 10.173.1.128 255.255.255.128dns-server 10.59.138.51default-router 10.173.1.129option 150 ip 10.59.138.51domain-name cisco.com!interface GigabitEthernet1/0! description Internal Interface to Session into Switchip address 1.1.1.1 255.255.255.0power inline never!interface GigabitEthernet1/0.193! description data vlan for Data only devicesencapsulation dot1Q 193ip address 10.173.193.1 255.255.255.128!interface GigabitEthernet1/0.200! description voice vlan for Voice only devicesencapsulation dot1Q 200ip address 10.173.1.129 255.255.255.128!•
interface FastEthernet1/0/3 - 1/0/21! description phone with pc connected to phoneswitchport access vlan 193switchport mode accessswitchport voice vlan 200!interface FastEthernet1/0/22! description just PC onlyswitchport access vlan 193switchport mode access!interface GigabitEthernet1/0/2! description internal trunk port to the access routerswitchport trunk encapsulation dot1qswitchport mode trunk!Dual-Tier Branch Profile
The dual-tier branch profile looks similar to the single-tier branch profile; however, the switch is external to the router.This profile allows higher port density than an integrated switch module.
The integrated 10/100/1000 interfaces on the access routers are used as Layer 3 trunks connecting four ports on the switch configured as trunks. There are two connections on each access router to the switch. One connection is dedicated for voice traffic and one is used for data traffic. The two connections also provide a layer of redundancy that is discussed in High Availability. In this design, the default gateways for voice and data reside on two different dot1q subinterfaces. There is no Layer 2 switch on the access routers; therefore, there are no loops in this topology as well.
As with the single-tier branch profile, two VLANs have been created: VLAN 101 for voice traffic and VLAN 102 for data traffic. Again, random ports have been configured for voice only, data only, and a data device connected to an IP phone port for configuration completeness. Following are the LAN configurations for the dual-tier profile:
•
ip dhcp pool data_lannetwork 10.173.111.0 255.255.255.0dns-server 10.102.6.247default-router 10.173.111.1!ip dhcp pool voice_lannetwork 10.173.110.0 255.255.255.0dns-server 10.59.138.4default-router 10.173.110.1option 150 ip 10.59.138.51domain-name cisco.com!interface GigabitEthernet0/1no ip addressduplex autospeed auto!interface GigabitEthernet0/1.100! description Data Traffic Ports - ACTIVEencapsulation dot1Q 102ip address 10.173.110.3 255.255.255.0standby 100 ip 10.173.110.1standby 100 priority 90standby 100 preempt!interface GigabitEthernet0/1.101! description Voice Traffic Ports - STANDBYencapsulation dot1Q 101ip address 10.173.111.3 255.255.255.0standby 101 ip 10.173.111.1standby 101 priority 120standby 101 preemptstandby 101 track Serial0/0/0.17 50!•
ip dhcp pool data_lannetwork 10.173.110.0 255.255.255.0dns-server 10.59.138.4default-router 10.173.110.1!ip dhcp pool voice_lannetwork 10.173.111.0 255.255.255.0dns-server 10.59.138.51option 150 ip 10.59.138.51default-router 10.173.111.1domain-name cisco.com!interface GigabitEthernet0/1no ip addressduplex autospeed auto!interface GigabitEthernet0/1.100! description Data Traffic Ports- STANDBYencapsulation dot1Q 102ip address 10.173.110.2 255.255.255.0standby 100 ip 10.173.110.1standby 100 priority 120standby 100 preemptstandby 100 track Serial0/0/0.1 50!interface GigabitEthernet0/1.101! description Voice Traffic Ports -ACTIVEencapsulation dot1Q 101ip address 10.173.111.2 255.255.255.0standby 101 ip 10.173.111.1standby 101 preempt!•
interface GigabitEthernet1/0/2! description Trunk port to Access Router 1switchport trunk encapsulation dot1qswitchport mode trunk!interface GigabitEthernet1/0/3! description Trunk port to Access Router 2switchport trunk encapsulation dot1qswitchport mode trunk!interface GigabitEthernet1/0/4! description Trunk port to Access Router 1switchport trunk encapsulation dot1qswitchport mode trunk!interface GigabitEthernet1/0/5! description Trunk port to Access Router 2switchport trunk encapsulation dot1qswitchport mode trunk!interface GigabitEthernet1/0/6 - interface GigabitEthernet1/0/23! description phone with PC connected to phoneswitchport access vlan 102switchport mode accessswitchport voice vlan 101!interface GigabitEthernet1/0/24- interface GigabitEthernet1/0/28! description data only portsswitchport access vlan 102switchport mode access!Multi-Tier Branch Profile
Of first concern when scaling a branch solution to accommodate the user population is often the available bandwidth compared to the size. How much bandwidth is required for 100 users? The access router must have sufficient CPU resources to support the branch users. The final consideration that is often overlooked is how to provide sufficient LAN ports for all end-users and devices at the branch. In branch offices of less than 46 users, the connectivity can be a single switch. In larger branch offices, several switches can be used. The connectivity between these switches and the router can become fairly complex because the requirements for high availability often leads to multiple Layer 2 paths.
When switches are connected with more than one Layer 2 connection between them, Spanning Tree (802.1d) must be considered. Spanning Tree prevents loops in redundant Layer 2 topologies. Spanning Tree neither converges as quickly as most routing protocols, nor is it as robust, and can be difficult to troubleshoot. Minimizing the complexities of the Spanning Tree configuration is therefore advantageous. One possibility is to minimize the broadcast domain enabling Layer 3 routing to the wiring closet (access-layer) switch. VLANs are limited to a single switch and extend only to the client NIC card. This approach is discussed as part of routing service for the multi-tier branch profile. Another option is to build a stack of switches. Stackwise is a method to physically attach the switches to a common 32 Gbps backplane cable. When a switch becomes a member of the stack, it logically appears as an physical extension of the stack master switch. Integrated switches such as the NME-XD-24ES-1S-P can be members of a stack. When an integrated switch is used, it should be the stack master. Only one integrated stackable switch can be u
