Re: BGP Mulithoming question [7:121]
Hi, I haven't the BSCN book, but using standard BGP terminology is easy to answer your question. Suppose you are multihomed to two providers' ASs, e.g. AS1 and AS2 1)Default Routes from All providers This means both AS1 and AS2 will send you a default route (prefix 0.0.0.0/0), and you will typically use one of the links as primary and the other as a backup; you won't have any information to help you choose which path is "best" for your outgoing traffic, so you'll need to tweak local_pref to prefer the primary, and you will need to influence traffic coming into your AS by AS_PATH-prepending on your backup link or by using communities, if supported by your provider. 2)Customer Default Routes from AS providers This is almost the same as 1), but the provider which you intend to use as backup (e.g. AS2) will send you not only a default, but also the routes for its customers. This way, you can make better use of your backup link, by using it not only as a backup but also as a way of reaching quickly all AS2's customers, without going via AS1 and who-knows-what interconnections. As to the outbound traffic, the local_pref and AS_path setup may stay very similar to 1), since all AS2 customers will be reached by the backup links since there will be a specific route as opposed to a default; the inbound traffic configuration is slightly more tricky, and to solve it correctly you might need to use communities, or accept a small percentage of "primary" traffic coming in via the backup link. 3)Full Routes from All providers You get full routing tables from both providers - this way you are free to use both links in a way that minimizes AS_PATH length for all destinations. You usually do this if you are NOT using a primary/backup setup, but you "simply" want to interconnect to different provider, chose the best path to any destination and make sure that traffic can reach your network in the best way. You will also need bog boxes (not 2511)!!! Please check Halabi's BGP book for examples and case studies! Hope this helps, Saverio Message Posted at: http://www.groupstudy.com/form/read.php?f=7i=145t=121 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: BGP Mulithoming question [7:121]
On Tue, 10 Apr 2001, Larry Lamb wrote: 2)Customer Default Routes from AS providers You receive 0.0.0.0 routes and all other routes that originate from the providers AS (providers internal routes). This will route traffic coming from customers of this provider back to the provider. All other traffic will get balanced out both providers. You usually don't balance traffic in this situation - just use the higher-speed/higher-reliability link as primary, and the other as backup. This is because connectivity between different providers can vary a lot, and you have no way of deciding which provider would be best for a certain destination. That is why you need to use LOCAL_PREF, AS_PATH prepend or communities... 3)Full Routes from All providers Here you have all known routes being send to you. This provides the best routing outbound for customer connections, but requires a great deal of memory because of the number of routes (75,000 +) in the tables. Actually, we're just under 100.000 routes right now! Check this: http://www.employees.org/~tbates/cidr-report.html#General_Status Cheers, Saverio Message Posted at: http://www.groupstudy.com/form/read.php?f=7i=147t=121 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: Alternative to Prepanding [7:150]
On Mon, 9 Apr 2001, Howard C. Berkowitz wrote: Also, I'm not sure what you mean by "community." If you refer to the global routing system, you can't force it to do anything. If you are referring to the BGP community attribute, it doesn't inherently have any conditional control. I think I understand the source of the confusion.. Check RFC1998.. almost all big providers support it! Cheers, Saverio Message Posted at: http://www.groupstudy.com/form/read.php?f=7i=150t=150 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: bgp path selection criteria
On Tue, 21 Nov 2000, Yee, Jason wrote: Anyone here knows which BGP path criteria takes precedence ? AS_PATH or local preference Local preference! See Halabi's book P.168 From what I read it is local preference , but in actual fact it is not so , why I said this is because I have a customer who prepends their prefixes many times then advertise them to us but on our side we set local preference to customers' routes to 90 which in fact will always come back to us if we do this but this is not happening Instead the prefixes go to another providers' link because their AS-PATH is shorter Here you lost me; what does it mean "prefixes goes to another provider's link"? Please consider that the higher the local preference, the "more preferred" is the route; can you please send to the list the output of "sh ip bgp prefix for one of the prefixes you are having problem with ? The output should be something like this: (fake example ) 3300 3300 3300 3300 8933 137, (aggregated by 137 193.206.129.254), (received used) 195.206.65.137 (metric 6) from 195.158.226.160 (192.121.158.8) Origin IGP, localpref 134, valid, internal, atomic-aggregate, best Community: 1755:80 1755:666 1755:1000 1755:2000 Originator: 192.121.158.8, Cluster list: 195.158.226.160 3300 3300 3300 3300 8933 137, (aggregated by 137 193.206.129.254), (received used) 195.206.65.137 (metric 6) from 195.158.226.161 (192.121.158.8) Origin IGP, localpref 104, valid, internal, atomic-aggregate Community: 1755:80 1755:666 1755:1000 1755:2000 Originator: 192.121.158.8, Cluster list: 195.158.226.161 In this example the first route is chosen for its highest local preference.. Can you post an example of your output??? Cheers, Saverio _ FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: about OSPF,
On Wed, 11 Oct 2000, Brian wrote: On Thu, 12 Oct 2000, gary wrote: hi guys: i have 2 question: (1)is the split horizon avaiable in OSPF , yes Hi, If I remember correctly, split horizon is needed (and applicable) only on distance vector (RIP, (E)IGRP).. link state routing protocols like OSPF do not need it. Cheers, Saverio _ FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: about OSPF,
On Thu, 12 Oct 2000 [EMAIL PROTECTED] wrote: Actually, EIGRP is a hybrid protocol which means it is distance vector and link-state. Split horizon kills routing loops which can occur in either of the 2 so it is indeed available in OSPF. I know this for a fact because there is a command to disable it in OSPF: no split-horizon ospf, or something like that... Hi, Thanks for correcting me on EIGRP (I really don't know much over it). As to the OSPF split-horizon feature, can you point me to some information about what it does and why it is needed ? As far as I understand every OSPF router builds its own link state database, so there would be no meaning in using split horizon as defined in http://www.cis.ohio-state.edu/htbin/rfc/rfc2453.html paragraph 3.4.3. Is this some Cisco proprietary extension/feature ?? Many thanks, Saverio _ FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: Token Ring
On Sun, 8 Oct 2000, FRS wrote: Hi everyone, Just messing around with a router did a show int token 0 command says: Interface up, Line protocol down. When you issue the no shut command on the interface it still says Interface up, Line protocol down. Why is this? 'Cause nothing is physically connected to the interface, or, if it is, it is not working properly! Cheers, Saverio **NOTE: New CCNA/CCDA List has been formed. For more information go to http://www.groupstudy.com/list/Associates.html _ UPDATED Posting Guidelines: http://www.groupstudy.com/list/guide.html FAQ, list archives, and subscription info: http://www.groupstudy.com Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: Questions about Spanning Tree!
On Thu, 28 Sep 2000, Gem wrote: I have two questions about Spanning tree algorithm, 1. How to transmit BPDUs? use Broadcast or Multicast or Unicast? 2. Are Root bridge and Designated bridge the same or different one(s) in multiple bridges LAN? Hello Gem, 1) They transmit using multicast address addigned to "all bridges on the lan" on multicast-enabled LANs; for 802.5 a special functional address is used.. 2) Oops.. the root bridge is only one (like the Highlander) in the spanning tree; but there is a designated bridge for every lan in the spanning tree.. I recommend you to read Radia Perlman's book (Interconnections), it is relly good.. Hope this helps, Saverio **NOTE: New CCNA/CCDA List has been formed. For more information go to http://www.groupstudy.com/list/Associates.html _ UPDATED Posting Guidelines: http://www.groupstudy.com/list/guide.html FAQ, list archives, and subscription info: http://www.groupstudy.com Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
Re: Fw: Questions about Spanning Tree!
On Thu, 28 Sep 2000, WANG wrote: I have both Clark Hamilton and Radia books on hand. Not finding how BPDU exchanges between Bridges. Perlman book, 1st edition, paragraph 3.3.1, pag 55, last 3 lines ;) A bridging domain has one and only one *Root* bridge. But if there are lots bridges in a lan (part of the bridging domain), and the *Root* bridge is in this lan, it should elect a *Designated* bridge for this lan. What's the relation between the two bridges? The root bridge is also the designated bridge for each lan connected to it, since the designated bridge is defined as the bridge which is closest to the root bridge in a given LAN. (does this does make sense!?!?!?). This is a little more complex in case multiple ports of the root brisge are connected to the same lan... but that's a pathological case: instead of "designated bridges" you can think of "designated ports".. Finished 350-001, I'll continue 350-014 next week, these two problems bother me. Hope you can sleep better now !!! Good luck, Saverio **NOTE: New CCNA/CCDA List has been formed. For more information go to http://www.groupstudy.com/list/Associates.html _ UPDATED Posting Guidelines: http://www.groupstudy.com/list/guide.html FAQ, list archives, and subscription info: http://www.groupstudy.com Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]