Connectivity Guide

Multihomed AS Maintains connections to more than one other AS. This group allows the AS to remain connected to the Internet if
a complete failure occurs to one of their connections. This type of AS does not allow trac from one AS to pass
through on its way to another AS.
Stub AS Connected to only one AS.
Transit AS Provides connections through itself to separate networks. For example, Router 1 uses Router 2—the transit AS, to
connect to Router 4. Internet service providers (ISPs) are always a transit AS because they provide connections
from one network to another. An ISP uses a transit AS to sell transit service to a customer network.
When BGP operates inside an AS - AS1 or AS2, it functions as an Internal Border Gateway Protocol (IBGP). When BGP operates between
AS endpoints - AS1 and AS2, it functions as an External Border Gateway Protocol (EBGP). IBGP provides routers inside the AS with the
path to reach a router external to the AS. EBGP routers exchange information with other EBGP routers and IBGP routers to maintain
connectivity and accessibility.
Classless interdomain routing
BGPv4 supports classless interdomain routing (CIDR) with aggregate routes and AS paths. CIDR denes a network using a prex consisting
of an IP address and mask, resulting in ecient use of the IPv4 address space. Using aggregate routes reduces the size of routing tables.
Path-vector routing
BGP uses a path-vector protocol that maintains dynamically updated path information. Path information updates which returns to the
originating node are detected and discarded. BGP does not use a traditional Internal Gateway Protocol (IGP) matrix but makes routing
decisions based on path, network policies, and/or rule sets.
Full-mesh topology
In an AS, a BGP network must be in full mesh for routes received from an internal BGP peer to send to another IBGP peer. Each BGP
router talks to all other BGP routers in a session. For example, in an AS with four BGP routers, each router has three peers; in an AS with
six routers, each router has ve peers.
Sessions and peers
A BGP session starts with two routers communicating using the BGP. The two end-points of the session are called peers. A peer is also
called a neighbor. Events and timers determine the information exchange between peers. BGP focuses on trac routing policies.
Sessions
In operations with other BGP peers, a BGP process uses a simple nite state machine consisting of six states—Idle, Connect, Active,
OpenSent, OpenConfirm, and Established. For each peer-to-peer session, a BGP implementation tracks the state of the session.
The BGP denes the messages that each peer exchanges to change the session from one state to another.
322
Layer 3