Re: OSPF aggreation question [7:26091]
Ok Here it is, this is how we tackled the OSPF Aggregation problem. To properly aggregate addresses within an internal type OSPF advertisement you must use the Area x range statement, this would not work because of the multi area setup and non-contiguous addresses. Stabilization of area 0 was a must because of the service provider environment, as well Link state changes would be constant due to customer networks. This would prevent us from using standard type areas due to routing table size and link state flapping. We needed to reduce the size of area 0 with summarization on the ABR's and keep all routes local to the areas to reduce the overall size of the routing table. We used NSSA areas because of one major feature type 7 LSA's. What this allowed us to do is convert all routes to external type routes and aggregate the mutli areas on the ABR's. External type routes can be summarized using a summary statement with the summary being non area specific. To do this within an NSSA area you simply add the statement redistribute connected with some sort of filter like a prefix list and route map combination to prevent un-wanted routes, If you have any static routes you redistribute the static routes with the redistribute static statement with the same sort of filter control. This will convert all the area routes to type 7 LSA's (which are external) and allow you to summarize the routes with the summary-address command. Cool Huh ... Mark Message Posted at: http://www.groupstudy.com/form/read.php?f=7i=26345t=26091 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: OSPF aggreation question [7:26091]
7500's can handle it...I would want RSP4's at a minimum. 200 routers can be in a full mesh and a single IGP area if you wanted. You probably only want to mesh your core and depending on the pysical configuration use RRs for the access layer...Unless all your access routers are on the same segment I wouldn't expect any one router to have a large number of peers as the clients would only be peering with the RRs. A RR client can also be a RR for other clients as well... On Tue, 13 Nov 2001, Mark Paterson wrote: Excellent suggestion, That is our end goal. And considering we are running a number of 12012's that can handle lots of peering sessions, BPG would be an excellent solution. However if you had smaller routers at the core, say 7500's could this still work. The network has 200 access routers, that is quite a few peering sessions. Even with route reflectors. So far there are some great suggestions Thanks. Mark Message Posted at: http://www.groupstudy.com/form/read.php?f=7i=26206t=26091 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: OSPF aggreation question [7:26091]
Good old-fashioned re-addressing. Once you re-address, summarize at the abr's. Mark Paterson wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... All, I have a question that may have several answers, I have tried a few with varying results and would just like to see if anyone else has any suggestions. We run a Large Telco Data Backbone, most of which has been run on OSPF and BGP. Our distribution and access layers contain layer three functionality. The first customer data MAN rings that were built used EIGRP, we now need to standardize these old network segments and migrate them to OSPF. This is not the hard part, what is difficult is reducing the size of Area 0 so that it remains stable. Summarization is the only way. However the ip addresses used were not continuous and the routed ring is far to large to use one area. How would you successfully summarize the address over multi-areas and reduce the size of the area 0 table. Mark I Will post what we did later. Message Posted at: http://www.groupstudy.com/form/read.php?f=7i=26151t=26091 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: OSPF aggreation question [7:26091]
Can't re-address, these a public routable addresses, remember this is a Telco Service provider, If it where that easy we would have done it that way, this network also makes an enterprise look small by comparison. You need to really think about this one. Message Posted at: http://www.groupstudy.com/form/read.php?f=7i=26154t=26091 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: OSPF aggreation question [7:26091]
You have to readdress. ATT does it, so can you. I never said it was going to be fun. Mark Paterson wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... Can't re-address, these a public routable addresses, remember this is a Telco Service provider, If it where that easy we would have done it that way, this network also makes an enterprise look small by comparison. You need to really think about this one. Message Posted at: http://www.groupstudy.com/form/read.php?f=7i=26155t=26091 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: OSPF aggreation question [7:26091]
Use IBGP for customer routes and OSPF only for reachability (loopbacks and links between routers). Your primary concern with OSPF at that point will be the number of routers in a given area and link instability. If you have enough routers that summarization is still reqired. it is much easier to renumber the links between routers than customers. On Tue, 13 Nov 2001, Mark Paterson wrote: Can't re-address, these a public routable addresses, remember this is a Telco Service provider, If it where that easy we would have done it that way, this network also makes an enterprise look small by comparison. You need to really think about this one. Message Posted at: http://www.groupstudy.com/form/read.php?f=7i=26160t=26091 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: OSPF aggreation question [7:26091]
Why go away from EIGRP? Are you putting in No'tell or something? (WARNING! 2 cents: If aint broke.) It seems that you may be able to place some those discontiguous nets into some stub or NSSA's. My gut thought is to limit the core router and one interface hop as part of Area 0 and then begin to creat your areas and summarize. This is just a guess based on limited information and no drawings to look at. But I could be wrong - Dennis Miller Best Regards, Chris Watson, CCNP Mark Paterson wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... All, I have a question that may have several answers, I have tried a few with varying results and would just like to see if anyone else has any suggestions. We run a Large Telco Data Backbone, most of which has been run on OSPF and BGP. Our distribution and access layers contain layer three functionality. The first customer data MAN rings that were built used EIGRP, we now need to standardize these old network segments and migrate them to OSPF. This is not the hard part, what is difficult is reducing the size of Area 0 so that it remains stable. Summarization is the only way. However the ip addresses used were not continuous and the routed ring is far to large to use one area. How would you successfully summarize the address over multi-areas and reduce the size of the area 0 table. Mark I Will post what we did later. Message Posted at: http://www.groupstudy.com/form/read.php?f=7i=26162t=26091 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: OSPF aggreation question [7:26091]
In a message dated 11/13/01 4:57:36 PM Central Standard Time, [EMAIL PROTECTED] writes: Subj: Re: OSPF aggreation question [7:26091] Date: 11/13/01 4:57:36 PM Central Standard Time From: [EMAIL PROTECTED] (Mark Paterson) Sender:[EMAIL PROTECTED] Reply-to: [EMAIL PROTECTED] (Mark Paterson) To:[EMAIL PROTECTED] Mark, Maybe a NAT solution, although it would be logistically complicated and would have to be lab tested. I work for an enterprise and we recently used something similar on a migration project. NDA prevents me from giving any details. :) My .02c, ROb H. NP, DP, blah,blah,blah. Can't re-address, these a public routable addresses, remember this is a Telco Service provider, If it where that easy we would have done it that way, this network also makes an enterprise look small by comparison. You need to really think about this one. Message Posted at: http://www.groupstudy.com/form/read.php?f=7i=26167t=26091 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: OSPF aggreation question [7:26091]
Excellent suggestion, That is our end goal. And considering we are running a number of 12012's that can handle lots of peering sessions, BPG would be an excellent solution. However if you had smaller routers at the core, say 7500's could this still work. The network has 200 access routers, that is quite a few peering sessions. Even with route reflectors. So far there are some great suggestions Thanks. Mark Message Posted at: http://www.groupstudy.com/form/read.php?f=7i=26182t=26091 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: OSPF aggreation question [7:26091]
Your a funny guy :-) come on are you not going for your CCIE, There is a way of doing this with out BGP or Re numbering. Mark Message Posted at: http://www.groupstudy.com/form/read.php?f=7i=26183t=26091 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
