EIGRP Bounded Updates

Unanswered Question
Mar 5th, 2008

What do it mean that EIGRP features "Bounded Updates"?

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
lamav Wed, 03/05/2008 - 10:46


A bounded update is a triggered route update that is sent only to those routers that need the information. The purpose is to conserve bandwidth.



Richard Burts Wed, 03/05/2008 - 10:55


To supplement the answer from Victor it may help to understand "routers that need the information". The thing that makes the update bounded is usually summarization or perhaps route filtering. The update is propagated to all neighbors who should know that particular route. But at the point where that route gets summarized or that route gets filtered out then there is no need to advertise that prefix any further.



lamav Wed, 03/05/2008 - 11:00


can you clairfy "who should know that particular route"?

Can you post a configuration example?



Richard Burts Wed, 03/05/2008 - 11:25


For example

interface Fast0/0

ip address

router eigrp 10


This will generate an advertisement for to all neighbors of that router (they should know this route). Those neighbors will advertise that route (prefix) to their neighbors (they should know this route). Now assume that one of these neighbors creates a summary address for on its interfaces. The neighbors out those interfaces have no need to know but instead will need to know

Alternatively perhaps some neighbor has configured a distribute list to deny out some interface. The neighbors out that interface do not need to know that route.

Does that help?



lamav Wed, 03/05/2008 - 11:36

This will generate an advertisement for to all neighbors of that router (they should know this route).

OK, that clears up that.

It's a language thing, you know...the words chosen...

As far as "need to know" routes, I guess it would be better stated as "allowed to know". IOW, the neighbors out that interface are only "allowed to know" the summary route.

I know it sounds like hair-splitting semantics, but in this case, a slight change in the words can confuse the issue considerably.



Richard Burts Wed, 03/05/2008 - 11:57


I absolutely agree with you that language and semantics are very important and frequently how we say things (the words that we choose) may help to clarify or may cause confusion.

I can see the alternative of describing it as "allowed to know" but I believe that there is a sense in which "need to know" is quite appropriate. In designing the network someone decides that certain neighbors need to know a certain level of detail (so that they will make appropriate forwarding decisions) and that other neighbors need to know less detail and will still make appropriate forwarding decisions.

[edit] As I think about it part of where I am coming from is an aspect of decision making theory that asks: how much information do you need to make the right decision. I believe that underlies the appropriate design of the network.



ruwhite Thu, 03/06/2008 - 05:45

Typically, we speak of bounded queries, not bounded updates.... Maybe that adds to the semantic muddle.



Mohamed Sobair Thu, 03/06/2008 - 06:26


The bounded update, is done when eigrp stablishes neighbor relationship, and whenever there is a changes on eigrp (periodic changes) and its a reliable.

the Query and the reply are also reliable handled by RTP.




This Discussion