BGP Convergence

[mvelazquezdsc]

Hi, i am having some problems in my metro network, bgp convergence i realy a problem, 180 seg is not an option, i was looking arround some fast convergence features like BGP Fast Peering Session Deactivation and BGP Next-Hop Address Tracking.

the core is based in 7600 and the U-PE are 3750ME. I have implemented this two features but nothing seems to be diferent. this is the config:

router bgp 222

no synchronization

bgp log-neighbor-changes

network 20.0.0.0

neighbor 1.1.1.1 remote-as 13591

neighbor 1.1.1.1 ebgp-multihop 3

neighbor 1.1.1.1 fall-over

neighbor 3.3.3.3 remote-as 13591

neighbor 3.3.3.3 ebgp-multihop 3

neighbor 3.3.3.3 fall-over

no auto-summary

the "fall-over" is only configured in one side.

Is there something missing? or am i using the wrong feature?

thanks a lot

Mario

[balajitvk]

Hi Mario,

It is the correct feature and relevant to improving convergence. But could u tell us wht u try to achieve and what u mentioned as 180sec and where u find this value.

Rgs

[mvelazquezdsc]

Hi

i was mentioning 180s as the dead timer for bgp peer, because for one of my customers needs a convergence time lower than 20s. this is the topology:

CE1---SW1----PE1----MPLS----PE2--------CE2

CE3---SW2----PE3 PE4--------CE4

CE1 and CE3 are in the same site the same for CE2 and CE4, the CE-PE connection is bgp.

the problem is between CE1/PE1 and CE3/PE3 becasuse when if one port fails in SW1/SW2 the bgp session will not be dropped until the dead timer expires.

How can this scenario be improved?

rdg.

[balajitvk]

Hi,

To improve this, u can reduce your holdtime interval so that broakage in the connectivity between CE1 and CE2 will be early understood and so that convergance will also happen early in the network.

U can use this command

Router(config-router)# timers bgp keepalive holdtime

or even u can apply per neighbor,

Router(config-router)# neighbor [ip-address | peer-group-name] timers keepalive holdtime

When a connection is started, BGP will negotiate the hold time with the neighbor. The smaller of the two hold times will be chosen. The keepalive timer is then set based on the negotiated hold time and the configured keepalive time.

Go thro.this link for better understanding,

http://www.cisco.com/en/US/products/ps6350/products_configuration_guide_chapter09186a00804435fc.html#wp1002274

Hope this will clear ur probs,

Rate if does,

Rgs

[mvelazquezdsc]

i just tried the change in the timers and works fine in lab, but when i configured the timers in the real network the cpu went up to 99%, i forgot to mention that CE1 and CE2 (3750ME) have more than 6000 ospf routes, is there any other bgp feature that we can use?

thanks a lot for your help

Mario

[balajitvk]

Hi Mario,

Yes, it is really CPU intensive task and needs to be applied carefully after studying the particular n/w and itzz traffic.

Another way is to use less advertisement timers. By default the bgp router will update itzz internal neighbhors once per every 5 sec and external thro. 30 secs about a particular route if subsequent changes occurs. But there are commands to reduce this time so that it will update itzz neigbhor early so that convergence will happen early.

neighbor {ip-address | peer-group-name} advertisement-interval seconds

no neighbor {ip-address | peer-group-name} advertisement-interval seconds

But the unfortunate is mostly and all steps taken towards reducing the bgp convergence will require more CPU and memory of the routers. So we need to chose one among them as our first preference as according to our n/w

Hope this will clear ur doubt,

Rate if it does,

Rgs

[cisco_lad2004]

Hi Mario

If it is an option and CE H/W allows. U coudl configure an Etherchannel thru SW, then PE so if one port fails teh other still guarantees ur TCP session and BGP would not notice.

Of course, this will not protect you if ur SW fails.

with BGP, my advice is to provide a good L2 redundancy and dont rely on BGP itself.

HTH

Sam

[mheusinger]

Hi,

one other option would be Bidirectional Forwarding Detection (BFD), which could trigger BGP session tear down. There are UDP packets exchanged at a configurable interval between two routers to monitor connectivity. If a configurable number of packets are missing, then BGP can be triggered to reroute. For technical details have a look at "BFD for BGP" at

http://www.cisco.com/en/US/products/ps6441/products_configuration_guide_chapter09186a00805387d7.html#wp1117903

and "Bidirectional Forwarding Detection" at

http://www.cisco.com/en/US/products/ps6017/products_feature_guide09186a00803fbe87.html

Hope this helps! Please rate all posts.

Regards, Martin