[c-nsp] question about BFD

[Saku Ytti]

On (2006-01-04 17:26 -0800), Dan Martin wrote:

Hey,

> I'd be interested in hearing about any operational experience with BFD.
> I think unless its used wisely it could be the equivalent of putting a
> rubber band over your grip safety, a specifically good idea could turn
> into a generally bad one, depending on the circumstances.

 We've labbed it in 7600 and our experience were not good, timers
that were 50ms by 5 were too aggressive and even though we 
configured 'process-max-time 20' BFD still jumped on high BGP
utilization. TAC said that we just need to have less aggressive
timers and that there's nothing wrong with the implementation. 
 Personally, I don't think that 50ms by 5 is not aggressive
in BFD world and were indeed copied from CCO somewhere.


> -----Original Message-----
> From: cisco-nsp-bounces at puck.nether.net
> [mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of Bruce Pinsky
> Sent: Wednesday, January 04, 2006 6:52 PM
> To: Srividya Rao
> Cc: cisco-nsp at puck.nether.net
> Subject: Re: [c-nsp] question about BFD
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Srividya Rao wrote:
> > Hey all,
> > 
> >  
> > 
> > I have a few questions regarding BFD on Cisco:
> > 
> >  
> > 
> > -          Does BFD notify its clients (registered protocols like
> > OSPF/ISIS) when a link comes up? If it does, what value does that hold
> > in the protocol's context? Every protocol has its own hello mechanism
> > wherein it not only recognizes its neighbors, but also exchanges some
> > protocol-specific information. Hence, does BFD's notification add any
> > value?
> > 
> > -          Is it possible to promise a turnover time of 50ms (or say
> > even 100ms) for links on VLANs? This is because, in the case of VLANs,
> > BFD will have to depend on ARP to resolve the neighbors IP address and
> > hence BFD's timing is dependent on ARP's timing.
> > 
> > -          Similar to the above question, how does Cisco implement BFD
> > for ISIS? For an ISIS neighbor, ARP need not be necessarily resolved.
> > So, does BFD request ARP to resolve the neighbor's IP?
> > 
> >  
> > 
> 
> Pretty specific implementation questions....are you trying to insure
> compatibility with Cisco as you are developing Force 10's capability?
> 
> - --
> =========
> bep
> 
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.0 (MingW32)
> 
> iD8DBQFDvF+ME1XcgMgrtyYRAoZJAKDdORlWBgBUPpDyntVlN35LMRqXKwCfepYo
> EXCHysnDl88gdug62SrTgLg=
> =kyPh
> -----END PGP SIGNATURE-----
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/
> 
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/
> 

-- 
  ++ytti