BGP is another IP routing protocol that enables you to route IP. BGP is defined in RFCs 1771 and 1267. BGP has evolved from version 1 to version 4. The earlier versions are now obsolete, and the standard version is 4, or BGP4. BGP running between autonomous systems is referred to as external BGP, or EBGP, while running BGP inside an autonomous system it is referred to as internal BGP, or IBGP.
BGP is the only routing protocol that runs over TCP using port 179. A TCP connection is manually configured between routers and is referred to as a peer connection, or the BGP neighbor. After a TCP peer connection is established, the BGP routers exchange their full routing tables. Keep-alives are then sent to ensure that the session stays active, and only topology changes are sent in the future. BGP uses a separate table called a BGP table to maintain routing links. Only when an entry is in the BGP table along with other conditions discussed later in this chapter will a network be inserted into the IP routing table. BGP also supports CIDR, which allows the routing table to be reduced by using summarization. In summary, BGP:
► Serves as a path-vector protocol.
► Runs over a TCP connection using port 179.
► Routes between and inside autonomous systems. The AS number must be unique in the routing domain to stop routing loops. The InterNIC that handles IP addresses also handles AS for the Internet.
► Uses a complex routing decision algorithm based on the 10 conditions, which will be discussed later in this section on BGP.
Let's now cover the BGP attributes that are used to select the preferred path to a remote network.
Was this article helpful?