The IP Over NBMA (ION) Archive

[smirnow@fokus.gmd.de (Michael Smirnow)]

Masataka,

thanks for your comment:

> > 7. Scalability
> > 
> >      In large physical ATM network a single EARTH server could become a 
> >      bottleneck. In case of ATM Forum UNI 4.0 implementation this problem 
>        ^^^^^^^^^^
> >      could be readdressed to the ILMI which is supposed to support group 
> >      ATM addresses. That is, EARTH server may resolve IP multicast 
>                                ^^^^^^^^^^^^^^^^^^^^^^^^
> >      address to ATM group address while changes to the group within the 
> >      ATM network will be managed by ILMI.
> 
> As the EARTH server is the bottleneck, the EARTH server, the bottleneck,
> CAN NOT resolve IP multicast addresses.

You are right, however, I have two points against your comment:

1/. The draft assumes some _future_ implementation of EARTH server, say, 
EARTH+ working under UNI4.0. EARTH+ will work only with senders, roughly 
speaking, while 'changes to the group within the ATM network 
will be managed by ILMI.'
2/. The draft's target is ATM subnet (ATM LAN); to my understanding these
LANs currently are not so large and when they are made large (with, say, PVCs
through ATM public network, then they have _at least_ different management, and
therefore, do not comprise a single 'ATM cloud' as defined in the draft.

> 
> If MLIS is small enough and there is no bottleneck problem, a single
> ATMARP server can resolve IP unicast addresses for the entire MLIS.

Sorry, I don't understand this comment: the draft proposes no changes 
to IP _unicast_ address resolution (?).

Thanks,

Michael