Re: ISP BGP Propagation

That is correct, for internet feeds I would not lower the timers. Also you
might consider BGP dampening. The point was to explain why you might notice
delays. For a protocol to be reliable and stable on the internet delays are
required to ensure network stability when links flap, routes are removed,
etc. If you have a private BGP connection internal to your network then
modification of the timers might make sense. It depends on the scenario and
requirements.

Steve Lyons

On Fri, Aug 21, 2009 at 10:04 AM, ALL From_NJ <all.from.nj_at_gmail.com> wrote:

> Don't lower the timers for an internet feed ...
>
> Depending on the platform and the amount of activity your router has, you
> will be spiking your CPU frequently, unnecessarily, and in some cases
> causing a reboot. ;-(
>
> I would not normally suggest to lower the timers unless you performed some
> stress testing and also checked other apps such as mcast w/ PIM, igmp etc
> ..., monitoring etc ...
>
> Fast convergence can be achieved in other manners ... and for internal
> networks people do sometimes lower the timers.
>
> As mentioned earlier in this email, 2-3 minutes is not bad ... but if you
> are offering services, the bottleneck is going to be DNS propagation ... so
> this will need to be done in advance, else you will be waiting the 24 to 48
> hours for DNS convergence.
>
> HTH,
>
> Andrew Lee Lissitz
>
>
>
> 2009/8/21 WW)W WWWW W <eshedalonie_at_gmail.com>
>
>> from what I can suggest is to lower the timers of the bgp to hello 3 and
>> 20
>> seconds hold-time.
>> by default is 60 seconds hello 180 hold-time.
>>
>> and to look in http://www.ris.ripe.net/cgi-bin/lg/index.cgi
>>
>> for the show ip bgp command.
>>
>>
>> Blogs and organic groups at http://www.ccie.net
>>
>> _______________________________________________________________________
>> Subscription information may be found at:
>> http://www.groupstudy.com/list/CCIELab.html
>>
>>
>>
>>
>>
>>
>>
>>
>
>
> --
> Andrew Lee Lissitz
> all.from.nj_at_gmail.com

Blogs and organic groups at http://www.ccie.net
Received on Fri Aug 21 2009 - 16:45:52 ART