> I haven't thought deeply about this, and I'm not suggesting that we call
> them "Resource Distribution Networks" (yet!), but the bandying about of
> 'content' without a solid definition has disturbed me for a while.
I tend to think in terms of distributed systems that implement a service
which might or might not be available at multiple locations on the network.
The service implements one or more named resources each of which whose
function might or might not be to provide (static or dynamic) content.
For this and other reasons it's useful to separate the problem of how
one maps from a resource name to a location on the network, from the
problem of how one actually makes the service available at multiple
locations with a reasonable degree of consistency (for that service).
If the service more-or-less passively provides content, this is a very
different problem than if the service absorbs content, or if it involves
substantial interaction. But a passive content replication service
can still be a useful substrate in implementing a more dynamic service -
provided the services are layered correctly.
(we should avoid thinking of services as inhernetly being the sources
of content...they can as easily be sinks)
Keith
This archive was generated by hypermail 2b29 : Thu Nov 18 2004 - 11:21:29 MST