Re: Eigrp Query Scope

That definitely makes sense given redistribution.

On Sun, Aug 2, 2009 at 6:16 PM, Anantha Subramanian Natarajan <
anantha.natarajan_at_gravitant.com> wrote:

> Hi Petr,
>
> Awesome ..Thank you very much
>
> Regards
> Anantha Subramaian Natarajan
>
> On Sun, Aug 2, 2009 at 5:14 PM, Petr Lapukhov <petr_at_internetworkexpert.com
> > wrote:
>
>> Correction to statement (2) above - the "next-hop" may be missing the
>> exact network due to filtering, not summarization.
>>
>>
>> 2009/8/2 Petr Lapukhov <petr_at_internetworkexpert.com>:
>> > Hi Everyone,
>> >
>> > Here is a short list of the things that limit query scoping. In short,
>> > query is stopped every time the "exact" route is not found in the
>> > topology table. Here is a list of the things that may cause this:
>> >
>> > 1) Network summarization. The query will be stopped at the "next-hop"
>> > router, as that one will not have an exact match after summarization.
>> > 2) Distribute-list filtering. Queries are not affected by distribute
>> > lists, but will effectively get stopped if the "next-hop" router will
>> > not have the network in question after summarization.
>> > 3) Stub routers. Well-know scalability feature - stubs do not allow
>> > for transit routing and greatly reduce the amount of queries.
>> > 2) Different EIGRP ASes. Generally, queries stop at the AS boundaries.
>> > However, if redistribution is configured between the two ASes, things
>> > may become more complicated.
>> >
>> > Look at the following "diagram" with R1 being on the boundary b/w AS1
>> and AS2:
>> >
>> > AS1---R1---AS2
>> >
>> > Imagine R1 is doing mutual redistribution between AS1 and AS2 and a
>> > query from AS1 reaches R1. If R1 has no ways to reach the network in
>> > question, it will send a reply and remove the network from the routing
>> > table. This, in turn, will make AS2 process go active (since the
>> > prefix is redistributed) for the SAME network and make R1 initiate
>> > another queury (leaked query) into AS2. The queury will propage AS2
>> > using the regular rules. If somehow the network will be found, R1 will
>> > update the routing table and may re-inject the prefix into AS1.
>> >
>> > As for the new query scope in AS2, it may be one hop or more,
>> > depending on the filtering/summarization configuration and/or the
>> > number of mutual redistribution points between AS1 and AS2. In optimal
>> > case, if you are using route summariztion and dont have multiple
>> > redistribution points, the leaked query will travel just one hop deep
>> > into AS2.
>> >
>>
>
>

-- 
Regards,
Joe Astorino - CCIE #24347 R&S
Technical Instructor - IPexpert, Inc.
Cell: +1.586.212.6107
Fax: +1.810.454.0130
Mailto:  jastorino_at_ipexpert.com
Blogs and organic groups at http://www.ccie.net