[AG-TECH] AG 2.0 and static multicast addresses

Jennifer Teig von Hoffman jtvh at bu.edu
Tue Feb 4 15:52:29 CST 2003

Hi everybody, and especially Ivan :-) ,

Having had a bit of time to reflect on today's town hall, I'm quite 
worried about the fact that AG 2.0 will shift us to a system of 
exclusively dynamically-generated multicast addresses for venues (even 
allowing for some transition time and special transition venues).

At first this worried me purely in the context of the implications of 
nodes running 1.x not being able to co-exist in virtual venues with 
nodes running 2.x; if this incompatibility exists, it's going to be a 
nightmare for those of us planning even very small events, even if the 
transitional rooms Ivan spoke of today were available. If you were, say, 
planning a meeting among 5 sites, you'd need to either be sure that 
everybody was running the same version of the software, or be sure to 
reserve the "transitional" venue where everybody could co-exist. My 
hunch is that most events would take place in these transitional rooms, 
since most of us wouldn't be able to find the time to ask everybody what 
version of AGTk they were running.

But then, after some time worrying about that, I started worrying about 
ports and firewalls and such. I'm already aware of a couple AG nodes 
where they have to ask their firewall admins to unblock specific 
ports/addresses in order to participate in a given meeting; surely that 
sort of firewall reconfiguration won't happen on the fly along with the 
dynamic address allocation. So people at those nodes wouldn't be able to 
upgrade to 2.0 unless they could convince their network security staff 
to make some substantial changes.

And I'm guessing that if I (a relatively non-technical user) am coming 
up with these concerns, there are probably more big issues here too -- 
it's a major change in the underpinnings of the AG.

- Jennifer

More information about the ag-tech mailing list