Header
Home | Set as homepage | Add to favorites
  Search the Site     » Advanced Search
Sections
Syndication


Blogroll:

||||| ALL Cisco-Network ARTICLES |||||  
CCIE Journey,
The CCIE Journey,


BGP Operation

Nov 30,2008 by alperen

image

BGP Operation
BGP is known as the routing protocol of the Internet. BGP allows for the communication of
routing information between different autonomous systems spread throughout the world.
Figure 8.1 shows numerous autonomous systems, which will utilize BGP to share routing
information. There are two forms of BGP used: internal BGP (iBGP) and external BGP (eBGP).
All BGP speakers contained within the same autonomous system use internal BGP (iBGP) to
communicate with one another. There are a couple of important items to note about multiple
BGP speakers within the same autonomous system. First, all of the BGP speakers must peer with
one another. This means that you must configure a full mesh for iBGP to operate properly. This
doesn’t mean all devices must be connected to one another—just that all of the BGP speakers
must have layer 3 reachability. iBGP will utilize the TCP protocol to form the peering sessions
between the iBGP peers. There are ways to overcome the full mesh limitations, but they will be
explained in Chapter 9.
Another important characteristic of iBGP peers is that they will not advertise iBGP-learned
routes to one another. Then how do these internal routes get distributed within the AS, you ask?
By the IGP, of course. BGP is not meant to replace the IGP, and there is no comparison between
iBGP and an IGP, so don’t get caught by that common fallacy. When an iBGP speaker advertises
an eBGP-learned route to its iBGP peers, which should be all BGP speakers in its AS, there is no
need for them to advertise this route to other iBGP peers, because all of them have been informed
of the route already by the original iBGP peer. If any of these iBGP speakers also speak eBGP, it
is perfectly acceptable for them to advertise this iBGP-learned route to their eBGP peer.
External BGP (eBGP) is utilized between BGP speakers in different autonomous systems.
Like iBGP, eBGP peering sessions require the BGP speakers participating to have layer 3 connectivity
among themselves. TCP is then utilized by eBGP to form the peering sessions.
After forming peers, the BGP speakers use the peering information to create a loop-free map
of the autonomous systems involved. This is also known as a BGP tree.
256 times read

Related news

» Confederations
by alperen posted on Dec 01,2008
» Overcoming Scalability Limitations of iBGP
by alperen posted on Nov 30,2008
» BGP Terminology
by alperen posted on Nov 30,2008
» Route Reflection
by alperen posted on Nov 30,2008
» Stub AS
by alperen posted on Nov 30,2008
Did you enjoy this article?
(total 0 votes)

comment Comments (0 posted) 

More Top News
CCSP-Cisco Certified Security Professional
Most Popular
Most Commented
Featured Author