Jim Starkey <jas@...> wrote:

>At 09:52 AM 5/19/00 +1000, Jan Mikkelsen wrote:
>>>Do we agree so far? If so, I think we can duck the issues of
>>>broadcast technologies and reliable delivery.
>>
>>Removing the requirement for correctness always makes things a bit easier
>>(ie: no recovery), but I think broadcast is important. Using broadcast
>>instead of individual reliable streams makes recovery more important.
>
>Broadcast is a LAN solution. It doesn't work on the Internet,
>which, in my opinion, rules it out.

Not so: IP multicast works across router boundaries and over WAN links.
Ref. RFCs 1112, 2236, 2365. Also of interest is RFC 2357.

Of course, some network operators might filter out that kind of traffic, and
UDP to those hosts unreachable by IP multicast would be necessary. You
could even further fall back to TCP. Also see
http://www.cs.columbia.edu/~hgs/internet/mbone-faq.html for links to
operators prepared to carry multicast traffic.

Addressing is an issue, but I don't think that is insurmountable. Ignoring
the Internet, broadcast is a win for multisite organisations because WAN
link utilisation is O(1) rather than O(n). Of course, the same is true for
a single LAN, but bandwidth is cheaper there.

On the Internet it is even more important to use a multicast technology
because each packet can incur a charge under some tariffs. By ruling out
these technologies, you ensure worst case utilisation of server and and
network resources.

Jan Mikkelsen
janm@...