On Fri, 2010-01-15 at 09:00 -0500, Joel M. Halpern wrote: > I could not find the actual critique in this message. I found a > "signature.asc", but no other attachments, and the message body is as shown. > > Yours, > Joel
That's because I forgot to attach the file :) Sorry for that. Here it is. // Javier
Name-Based Sockets Critique Name-based sockets contribution to the routing scalability problem is providing end hosts with a network interface which makes the end-host address (locator) agnostic. The name abstraction allows the hosts to use any type of locator, independent of format or provider. The projected result is that no provider independent addresses will be required for multi-homing and mobility solutions. Deployment: The main incentives and drivers are geared towards the development of new applications and in the slower migration of old applications to the new API. Also, not all applications can be ported to a FQDN dependent infrastructure, e.g. DNS functions. This hurdle can be overcome, and may not be a definite obstacle for the transition of a whole domain, but it needs to be taken into account when striving for mobility/multi-homing of an entire site. The transition of functions on individual hosts may be trivial, either through upgrades/changes to the OS or as linked libraries, however, upgraded applications require support from all interacting hosts. This can still happen incrementally, but it does require the server in a client/server application to maintain both an upgraded as well as a non-upgraded function during the transition phase. Edge-networks: Name-based sockets provide a new networking-API which is not necessarily backwards compatible (there is a track that suggests allowing IPs as names). This puts in question to which extent a whole edge-site might accept PA addresses. Name-based sockets may make an individual client agnostic to the networking medium, be it PA/PI IP-addresses or in a the future an entierly different networking medium. However, an entire edge-network, with internal and external services will not be able to make a complete transition in the near future. In short, new services may be implemented using name-based sockets, old services may be ported, but for a complete edge-network transition backwards compatibility may be a hinder.
signature.asc
Description: This is a digitally signed message part
_______________________________________________ rrg mailing list rrg@irtf.org http://www.irtf.org/mailman/listinfo/rrg