> > BE traffic as it is thought today, should not need an e2e negotiation to
> > work correctly. However it could be subject to a peer negotiation
> > (adjacent providers).
> >
> > But, as we are trying to introduce the notion of e2e services with
> > performance guarantees, will BE survive in the future ?
>
> It is not only e2e services in strict sense. A Service Negotiation may
> lead to an allowed aggregate of CoS traffic at a particular access link,
> while the egress interfaces from the transport domain are unspecified.
> Such a negotiation may equally well apply to BE.
End-users, IMO, should think of a service as being e2e, even though
destinations are not specified. They should negotiate an e2e service but
providers could implement it (or part of it) using BE.
There is a distinction between a service (and its associated QoS
guarantees) an user/application sees and the mechanisms used to implement
it. An over-provisioned network without any qos mechanism may be used to
implement a premium service, since the provider is able to negotiate
services and to control offered qos guarantees.
Regards
Carlos
This archive was generated by hypermail 2b29 : Mon Nov 06 2000 - 12:15:19 CET