This chapter covers the following key topics:

•  Building Peer Sessions
A walk-through of the negotiation process between BGP and its neighbors.

•  Sources of Routing Updates
The source and method by which routes get injected into BGP has implications for the accuracy and stability of routing information.

•  Overlapping Protocols: Backdoors
When alternate routes into and out of a network are offered by overlapping protocols, a method of ranking them by preference is available.

•  The Routing Process Simplified
The decision model by which BGP receives, filters, selects for usage, and advertises routes, as a continuous process.

•  Controlling BGP Routes
At the core of BGP is a collection of attributes that administrators can apply to control routing according to their networks' needs.

•  Route Filtering and Attribute Manipulation
An example-oriented, step-by-step look at how BGP permits or denies routes, applies filters, and manipulates attributes to define the set of routing updates that enter and exit an AS.

•  BGP4 Aggregation
Several specific scenarios involving different aggregation choices and how BGP4 accommodates them.

Chapter 5
Tuning BGP Capabilities

Up to this point, this book has been concerned primarily with general definitions of interior and exterior gateway protocols and an overview of their respective and interconnected tasks. The Border Gateway Protocol was also presented from the technical perspective of its functional elements. With this chapter, you will begin to consider more practical implementation details for BGP as part of the overall design problem in building reliable Internet connectivity. This chapter examines specific attributes of BGP and how they are applied individually and together to address this design problem. Although the terminology, attributes, and details of this chapter are specific to BGP, the general concepts and problems raised are pertinent to routing architecture design, regardless of what specific protocols are being utilized.

Building Peer Sessions

The previous chapter began examining the process of BGP neighbor negotiation at a fairly technical level and with an emphasis on the formats of messages exchanged during negotiation. This chapter now expands the examination to consider additional subtleties of the negotiation process. In addition, distinctions between internal and external BGP, which have practical implications in building peer sessions, are introduced in this section.

Although BGP is meant to be used between autonomous systems to provide an interdomain loop-free topology, BGP can be used within an AS as a pipe between border routers running external BGP to other ASs. A neighbor connection, also called a peer connection, between two routers can be established within the same AS, in which case BGP is called internal BGP (IBGP). A peer connection can also be established between two routers in different ASs. BGP is then called external BGP (EBGP). Figure 5-1 contrasts these environments.

Figure 5-1  Internal and external BGP implementations.

Upon neighbor session establishment and during the OPEN message exchange negotiation, peer routers compare AS numbers and determine whether they are peers in the same AS or in different ASs. The difference between EBGP and IBGP manifests itself in how each peer would process the routing updates coming from the other peer and in the way different BGP attributes are carried on external versus internal links.

The neighbor negotiation process is mainly the same for internal and external neighbors as far as building the TCP connection at the transport level. It is essential to have IP connectivity between the two neighbors for the transport session to take place. IP connectivity has to be achieved via a protocol different from BGP; otherwise, the session will be in a race condition. An example of a race condition follows: neighbors can reach one another via some IGP, the BGP session gets established, and the BGP updates get exchanged. The IGP connection goes away for some reason, but still the BGP TCP session is up because neighbors can still reach each other via BGP. Eventually the session will go down because the BGP session cannot depend on BGP itself for neighbor reachability.

---

Troubleshooting:  
Verifying neighbor reachability for the TCP session to come up.

---

An Interior Gateway Protocol (IGP) or static route can be configured to achieve IP connectivity. In essence, a ping packet, containing a source IP address (the IP address of one BGP peer) and a destination IP address (the IP address of the second peer), must succeed for a transport session to initiate.