Re: ISP failures and site multihoming [Re: Enforcing unreachabilityof site local addresses]

[Masataka Ohta <mohta@necom830.hpcl.titech.ac.jp>]

Michel;

> > Kurt Erik Lindqvist wrote:
> > This I thought was more or less standard. I was talking about
> > less than 100ms convergence.
> 
> Dude, this requires a keepalive or hello at 10ms intervals and a 25~30
> ms rtt. You might need to talk to a guy named Albert Einstein; he wrote
> interesting RFCs about the speed of light that, as far as I know, have
> not been debunked yet.

A simple end to end solution is to send multiple copies of a data to
multiple pathes. Mission critical application over SONET is already
doing so and it is said that, even if a path fails, not even a single
bit is lost.

As for IP multihoming, it is already implemented with a VoIP protocol
and is working fine. Senders send multiple copies of a RTP data. At
the receivers, RTP takes care of duplicate packet deletion. Having a
few voice streams is not a problem at all, of course.

							Masataka Ohta