[e2e] Clarifying the End-to-End Principle

[Jon Crowcroft]

In message <3C83533F.A73B9275 at castify.net>, Arnaud Legout typed:

 >>[1] K. Leonard, Research Areas in Computer Communication, In Computer
 >>Communication Review, ACM SIGCOMM, volume 4, July 1974.

um, i think you have one end mixed up with the other - surely?
its : Kleinrock, L. :-)

meanwhile, i don't think the end2end model is fully expressed for the new cases we have
introduced of communicatiosn services such as group (any to many) and mobile ad hoc,
 where the _community_ if users is an "end", rather than a single point.

this means that the model needs updating and revising, since the services that one
builds between communities needs to provide, in the usual parsimonious systems manner,
the minimum set of intermediate node functions that is useful to a community.

we only have 1 example of this so far: IP itself is an overlay on a bit transport
service that  maximises connectivity. 

active networking research is motivated by the extreme case argument that arbirtrary
re-programming of intermediate nodes means that we can service any community
requirement, which, whiel true, is far far far from parsimonious. a minimal set of
thigns one might consider adding include:

protection services for flow aggregates (include both service protection such as
int/diff/and vpn service, but also security, integrity and other common services)

ip multicast in its original general service form

mobile ad hoc ip routing....

and any combination of the above


mechanisms for thse entail also understanding the virtual community costs and some
economic model (e.g. congestion price or similar, fairness enforcement or similar) which
admits of aggregates (spatially, temporally etc) needs evolving....this is HARD.
and not made out of IP as we know it, jim...

make it so
 

 cheers

   jon