All things related to multicast

[Guy Almes <>]

Russ,
  I agree.
  Just today I got a Help! email from a member campus where the
installation of a new improved router, combined with lack of familarity
of multicast configuration on the new router, combined to result in such
a backslding.

  Given the strong showing of multicast-based streaming this week on the
VIMM, this could be a real problem.

  Regards,
        -- Guy

Russ Hobby wrote:
> 
> A few months back at a CENIC TAC meeting I polled the member campuses about
> the state of multicast and found that several campus had backslid and had
> less multicast penetration then a couple of years ago.  Most of those that
> backslid indicated it was because that they installed new equipment
> (usually routers) and had not turned on multicast, either because they saw
> no demand to do so, or the new equipment was different to configure
> multicast and they had not figured it out yet (again demand would have made
> a difference).  A couple of other cases  multicast was turned off because
> people did start using it, but that broke unicast (or all networking).
> 
> The latter was the case here at UC Davis.  We have an ATM Backbone running
> LANE.  Muitcast goes thorough the Broadcast Unknown Servers, a limited
> resource.  So when multicast traffic got to the point where it was swamping
> the servers, netops just turned it all off and will only turn on multicast
> on special request to specific locations.  Fortunately the ATM network is
> scheduled to be replaced in a couple of months.
> 
> Russ