Routing Protocol Fundamentals
Routing occurs when a router or some other Layer 3 device (for example, a multilayer switch) makes a forwarding decision based on network address information (that is, Layer 3 information). A fundamental question, however, addressed throughout this book, is from where does the routing information originate?
A router could know how to reach a network by simply having one of its interfaces directly connect that network. Perhaps you statically configured a route, telling a router exactly how to reach a certain destination network. However, for large enterprises, the use of static routes does not scale well. Therefore, dynamic routing protocols are typically seen in larger networks (and many small networks, too). A dynamic routing protocol allows routers configured for that protocol to exchange route information and update that information based on changing network conditions.
The first topic in this section explores the role of routing in an enterprise network. Then some of the characteristics of routing protocols are presented, to help you decide which routing protocol to use in a specific environment and to help you better understand the nature of routing protocols you find already deployed in a network.
The Role of Routing in an Enterprise Network
An enterprise network typically interconnects multiple buildings, has connectivity to one or more remote offices, and has one or more connections to the Internet. Figure 1-1 identifies some of the architectural layers often found in an enterprise network design:
- Building Access: This layer is part of the Campus network and is used to provide user access to the network. Security (especially authentication) is important at this layer, to verify that a user should have access to the network. Layer 2 switching is typically used at this layer, in conjunction with VLANs.
- Building Distribution: This layer is part of the Campus network that aggregates building access switches. Multilayer switches are often used here.
- Campus Backbone: This layer is part of the Campus network and is concerned with the high-speed transfer of data through the network. High-end multilayer switches
- are often used here.
- Edge Distribution: This layer is part of the Campus network and serves as the ingress and egress point for all traffic into and out of the Campus network. Routers or multilayer switches are appropriate devices for this layer.
- Internet Gateways: This layer contains routers that connect the Campus network out to the Internet. Some enterprise networks have a single connection out to the Internet, while others have multiple connections out to one or more Internet Service Providers (ISP).
Figure 1-1 Typical Components of an Enterprise Network
WAN Aggregation: This layer contains routers that connect the Campus network out to remote offices. Enterprises use a variety of WAN technologies to connect to remote offices (for example, Multiprotocol Label Switching [MPLS]).
Routing protocols used within the Campus network and within the WAN aggregation layer are often versions of Routing Information Protocol (RIP), Open Shortest Path First (OSPF), or Enhanced Interior Gateway Routing Protocol (EIGRP). However, when connecting out to the Internet, Border Gateway Protocol (BGP) is usually the protocol of choice for enterprises having more than one Internet connection.
An emerging industry trend is to connect a campus to a remote office over the Internet, as opposed to using a traditional WAN technology. Of course, the Internet is considered an untrusted network, and traffic might need to traverse multiple routers on its way from the campus to a remote office. However, a technology called Virtual Private Networks (VPN) allows a logical connection to be securely set up across an Internet connection. Chapter 2 , “Remote Site Connectivity,” examines VPNs in more detail.
Routing Protocol Selection
As you read through this book, you will learn about the RIPv2, RIPng, OSPFv2, OSPFv3, EIGRP, BGP, and MP-BGP routing protocols. With all of these choices (and even more) available, a fundamental network design consideration becomes which routing protocol to use in your network. As you learn more about these routing protocols, keeping the following characteristics in mind can help you do a side-by-side comparison of protocols:
- Scalability
- Vendor interoperability
- IT staff’s familiarity with protocol
- Speed of convergence
- Capability to perform summarization
- Interior or exterior routing
- Type of routing protocol
This section of the chapter concludes by taking a closer look at each of these characteristics.
Scalability
How large is your network now, and how large is it likely to become? The answers to those questions can help determine which routing protocols not to use in your etwork. For example, while you could use statically configured routes in a network with just a couple of routers, such a routing solution does not scale well to dozens of routers.
While all the previously mentioned dynamic routing protocols are capable of supporting most medium-sized enterprise networks, you should be aware of any limitations. For example, all versions of RIP have a maximum hop count (that is, the maximum number of routers across which routing information can be exchanged) of 15 routers. BGP, on the other hand, is massively scalable. In fact, BGP is the primary routing protocol used on the Internet.
Vendor Interoperability
Will you be using all Cisco routers in your network, or will your Cisco routers need to interoperate with non-Cisco routers? A few years ago, the answer to this question could be a deal-breaker for using EIGRP, because EIGRP was a Cisco-proprietary routing protocol.
However, in early 2013, Cisco announced that it was releasing EIGRP to the Internet Engineering Task Force (IETF) standards body as an Informational RFC. As a result, any networking hardware vendor can use EIGRP on its hardware. If you are working in an environment with routers from multiple vendors, you should ensure that your Cisco router has an appropriate Cisco IOS feature set to support your desired routing protocol and that the third-party router(s) also support that routing protocol.
IT Staff’s Familiarity with Protocol
You and the IT staff at your company (or your customer’s company) might be much more familiar with one routing protocol than another. Choosing the routing protocol with which the IT staff is more familiar could reduce downtime (because of faster resolutions to troubleshooting issues). Also, if the IT staff is more familiar with the inner workings of one routing protocol, they would be more likely to take advantage of the protocol’s nontrivial features and tune the protocol’s parameters for better performance.
Speed of Convergence
A benefit of dynamic routing protocols over statically configured routes is the ability of a dynamic routing protocol to reroute around a network failure. For example, consider Figure 1-2 . Router R1’s routing protocol might have selected the path through Router R3 as the best route to reach the 192.168.1.0 /24 network connected to Router R4. However, imagine that a link failure occurred on the Fast Ethernet link between Routers R3 and R4. Router R1’s routing protocol should be able to reroute around the link failure by sending packets destined for the 192.168.1.0 /24 network through Router R2.
Figure 1-2 Routing Protocol Convergence
After this failover occurs, and the network reaches a steady-state condition (that is, the routing protocol is aware of current network conditions and forwards traffic based on those conditions), the network is said to be a converged network . The amount of time for the failover to occur is called the convergence time .
Some routing protocols have faster convergence times than others. RIP and BGP, for example, might take a few minutes to converge, depending on the network topology. By contrast, OSPF and EIGRP can converge in just a few seconds.
Capability to Perform Summarization
Large enterprise networks can have routing tables with many route entries. The more entries a router maintains in its routing table, the more router CPU resources are required support the ability to do network summarization, although the summarization options and how summarization is performed do differ.
Network summarization allows multiple routes to be summarized in a single route advertisement. Not only does summarization reduce the number of entries in a router’s routing table, but it also reduces the number of network advertisements that need to be sent. Figure 1-3 shows an example of route summarization. Specifically, Router R1 is summarizing the 10.0.0.0 /24, 10.0.1.0 /24, 10.0.2.0 /24, and 10.0.3.0 /24 networks into a single network advertisement of 10.0.0.0 /22. Notice that the first two octets (and therefore the first 16 bits) of all the networks are the same. Also, as shown in the figure, the first 6 bits in the third octet are the same for all the networks. Therefore, all the networks have the first 22 bits (that is, 16 bits in the first two octets plus 6 bits in the third octet) in common. By using those 22 bits and setting the remaining bits to 0s, you find the network address, 10.0.0.0 /22.
Figure 1-3 Network Summarization
Interior or Exterior Routing
An autonomous system (AS) is a network under a single administrative control. Your company’s network, as an example, might be in a single AS. When your company connects out to two different ISPs, they are each in their own AS. Figure 1-4 shows such a topology.
Figure 1-4 Interconnection of Autonomous Systems
In Figure 1-4 , Company A is represented with an AS number of 65000. ISP 1 is using an AS number of 65100, and ISP 2 has an AS number of 65200.
When selecting a routing protocol, you need to determine where the protocol will run. Will it run within an autonomous system or between autonomous systems? The answer to that question determines whether you need an interior gateway protocol (IGP) or an exterior gateway protocol (EGP) :
- IGP: An IGP exchanges routes between routers in a single AS. Common IGPs include OSPF and EIGRP. Although less popular, RIP and IS-IS are also considered IGPs. Also, be aware that BGP is used as an EGP; however, you can use interior BGP (iBGP) within an AS.
- EGP: Today, the only EGP in use is BGP. However, from a historical perspective, be aware that there was once another EGP, which was actually named Exterior Gateway Protocol (EGP).
Routing Protocol Categories
Another way to categorize a routing protocol is based on how it receives, advertises, and stores routing information. The three fundamental approaches are distance-vector, linkstate, and path-vector.
