Re: IS-IS Verification [7:30957]

From: Howard C. Berkowitz (hcb@xxxxxxxxx)
Date: Fri Jan 04 2002 - 16:08:39 GMT-3


   
Found my problem that was causing inadvertent crossposting.
Hopefully, I haven't confused everyone on both general and ccielab.

At 12:57 PM -0500 1/4/02, Peter van Oene wrote:
>Date: Fri, 4 Jan 2002 12:57:43 -0500
>From: "Peter van Oene" <pvo@usermail.com>
>X-GroupStudy-Version: 3.1.1a
>X-GroupStudy: Network Technical
>To: cisco@groupstudy.com
>Subject: Re: IS-IS Verification [7:30957]
>Sender: nobody@groupstudy.com
>Reply-To: "Peter van Oene" <pvo@usermail.com>
>Precedence: bulk
>
>comments inline
>
>At 12:07 PM 1/4/2002 -0500, Howard C. Berkowitz wrote:
>> >Multi-area ISIS is really a CLNS topic. It is indeed not recommend
>> >that one run multi-area for IP only routing. Some use in transition
>> >networks may take place, but your better to simply interconnect with
>> >a proper L2 backbone than merge areas.
>>
>>First, let me be sure we are talking about the same thing. I think of
>>multi-area as implying a backbone. Are you using it in the sense of
>>multiple area identifiers in a level 1 area?
>
>I will agree that the more recent use of the term multi-area implies a ISIS
>domain with multiple areas connected by a backbone. In this case the
>discussion is about have a single IS participate in multiple L1 areas (ie
>multiple area ID's as you suggest)
>
>> >
>> >The command clns routing is only required if you intend to route
>> >clns. For IP only environments, it is not needed. In IP only, you
>> >configure one NET per ISIS instance, and one ISIS instance per L1
>> >area you intend to form adjacencies within. The configs your
>> >looking at on CCO likely include the command clns router isis on
>> >each interface. I doubt there are many IP only multi-area configs
>> >on CCO given it is not a recommended practise.
>>
>>The main application I've seen, for which there are some CCO
>>documents, are telco applications when you are merging the CMIP
>>infrastructure for several central offices or SONET control points,
>>in order to reduce the requirement for L2 routers.
>
>Henk Smit indicates in a NANOG presentation that at times, some telco
>resident ADM's may have had area sizing limitations thereby necessitating
>the use of numerous small areas. Without multi-area support, a similar
>number of routers would be required to properly interconnect these
>areas. I expect He and You are describing a similar usage.
>
>>Of course, CCIE lab usage and best current practices often are
>>orthogonal at best, and in different space-time continuua at worst.
>>:-)
>>
>> >
>> >Pete
>> >
>> >
>> >At 09:32 AM 1/4/2002 -0600, Michael C. Popovich wrote:
>> >>I have double checked on Documentation CD regarding IS-IS and the need
>> >>for CLNS. Peter is definitely right in the fact that CLNS is not
>> >>required to route IP using IS-IS. This question is for Peter and anyone
>> >>else with a good understanding of IS-IS.
>> >>
>> >>Reading through the Documentation CD on Integrated IS-IS I have a
>> >>question regarding multi-area. The way I read it I am assuming that if
>> >>you implement a multi-area IS-IS solution then you in fact do need to
>> >>enable CLNS routing. Is this true? The examples given in the
>> >>Documentation CD show that a single area IS-IS network doesn't need CLNS
>> >>routing but in the multi-area scenario they have CLNS routing enabled.
>> >>This isn't specifically addressed but I assume with the need for
>> >>different areas using different NET addresses requires the CLNS
>> >>protocol.
>> >>
>> >>Please correct me if I have misinterpreted this.
>> >>
>> >>Thanks,
>> >>
>> >>Michael Popovich



This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:56:16 GMT-3