The IP over ATM Mailing List Archive by date

[Grenville Armitage <gja@bellcore.com>]

	[..]
>>> Interestingly, ipmc-12 also currently mandates that there
>>> shall be _no more than one MCS_ per group. 

I've been prompted privately to refine this comment (aka I'm
not strictly correct in what I just said). ipmc-12 states
that in the absence of a multi-MCS synchronization protocol
there should only be one MCS per group. Ooops.

>>Is this relevant to the MARS<->client and/or MCS<->client interfaces?
>>Or is it just MARS<->MCS? (I guess I'd better go figure that one
>>out myself).

Indeed. Its primarily just MARS<->MCS interface. If you recall
during 1995 the intent for MARSv1 was for clients to be completely
unaware of MCSs, so there _is no_ MCS<->client interface. As far
as the average MARS Client is concerned the MCS is just a leaf
node that the MARS told it to use.

gja