hey again,
there is also this protocol being developed by the working groups which may be
of vital interest
http://www.web3d.org/WorkingGroups/vrtp/
the virtual reality transfer protocol
On Tue, 14 Nov 2000 11:28:29 -0500, John Stewart said:
> Hi Miriam;
>
> > Hmmm... does multicast work over ordinary internet connections -- dial-ups
> > that most people have? I had a funny feeling that it required connection to
> > the backbone (I am actually not totally sure what that even is :)
>
> Well, there are 2 ways of answering this, so I'll answer both.
>
> 1) Multicast is "being deployed". It is getting better and better; and
> protocols
> exist that allow one to multicast over dialup (a simplification...
> but...)
>
> 2) For one of our projects, when we had multicast problems over the
> atlantic,
> a German colleague set up what is called a "packet reflector", which is
> kind
> of a repeater station; you point yourself at this, and any traffic that
> anyone sends gets repeated to everyone else. It actually worked very
> well,
> much better than multicast across the atlantic.
>
> This repeater can be part of the multicast group, and just send/receive
> packets to non-multicast enabled hosts. The repeater also is different
> than a server, in that there is no state information stored. It is
> simply
> a "send and forget" type of service.
>
> Does this help?
>
> Also, as Dr. Frog states in another email message, there is also a
> project
> in France that multicasts virtual worlds, and I have heard of others,
> too.
> I guess that I am partial to VNet, and thus have the desire to take the
> VIP
> and multicast that. :-)
>
> John.
>
>
> > If it does allow connection between common dial-ups then this is probably
> > the way we should be pushing. From my understanding of multicast it uses a
> > packet that isn't addressed to one individual, but to a group of people, so
> > each "server" only has to send one packet to all people receiving. And all
> > people "subscribing" to that address receive it. This is a great saving for
> > each machine.
>
>
