• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 4923
  • Last Modified:

BGP main loop prevention mechanism

I have read this :

((Since BGP does not use a discernible metric value as IGPs do, the main loop prevention mechanism built into iBGP is that fact that routes learned from an iBGP neighbor cannot be advertised onto another iBGP neighbor. Therefore,
this stipulation implies that all iBGP speaking devices must establish direct peering relationships with all other iBGP devices within your autonomous system)).

I am aware that we have to have full mesh peering relations within IBGP, that means, any IBGP speaker will not pass route update to other IBGP speaker unless it has peer relationship with it . In this way, we are preventing loops from happening .

But I could not get what he mean by his comparison when he says :

 “Since BGP does not use a discernible metric value as IGPs do”
0
zillah
Asked:
zillah
1 Solution
 
neowolf219Commented:
Hey zillah,

Okay, keep in mind that when doing any type of route engineering with IGPs, as long as the administrative distance is the same (ie, OSPF = 110, RIP = 120, etc, OSPF is always preferred over RIP regardless of metric), the metric will be the determining factor.  This isn't so in BGP.

The "loop" prevention techniques in iBGP is saying that I cannot advertise the route I received back out that interface.  To affect how traffic moves through an AS to get to another AS, many attributes can be set, such as Weight, Local Preference (to affect how traffic leaves your AS) and MED and AS-PATH affects how traffic comes into your AS).  

Notice how I'm saying "into" and "leaving" your AS.  If you are originating a route within your AS, then since everything is fully meshed the router will be the next-hop for that advertised BGP route within your iBGP (remember, fully meshed).  Hence why BGP doesn't use a discernible metric.  

In IGP, you might not be fully meshed, so you will prefer the path with the lowest metric, hence why it is important (most IGPs aren't fully meshed).  But remember with IGPs, administrative distance first, then metric.

HTH
0
 
pjtemplinCommented:
Wrong, iBGP can always advertise a route back out the interface it was received on.  There is no such limitation in BGP.

BGP attributes, assuming no route map processing, won't change as a route is passed around an AS.  That's the "no discernable metric" part of that statement.  Therefore, since nothing increments, if BGP routers passed iBGP-learned routes to other iBGP neighbors, there'd be no way to tell it's looping through the AS.  As a result, routers don't pass iBGP-learned routes to other iBGP neighbors, unless an iBGP mesh scaling trick is used.
0

Featured Post

Free Tool: SSL Checker

Scans your site and returns information about your SSL implementation and certificate. Helpful for debugging and validating your SSL configuration.

One of a set of tools we are providing to everyone as a way of saying thank you for being a part of the community.

Tackle projects and never again get stuck behind a technical roadblock.
Join Now