RE: [RRG] Are host-stack modifications allowed or disallowed ?

[Randall Atkinson <rja@extremenetworks.com>]

Earlier, Sheng Jiang wrote:
% It is obvious for everyone, that modifications on both hosts and routers
% are more expensive than modifications on only routers.

Agreed.

However, at least one concept, that of a true ID/Locator split,
requires no router changes at all.

An approach with a stack change to hosts, but no router changes
has different tradeoffs from an approach with router changes
but no host stack changes:

(off the top of my head; no doubt I've missed various things)

A) Host changes

- The number of hosts is much greater than the number of routers.
- Decision matrix for deployment is potentially larger (more host OSs
  than router OSs, so more implementers need to decide to change things)
+ Feature can be deployed whether or not ISPs decide they want
   to make changes.
+ The Internet potentially scales better because state in the core of
   the network is reduced.
+ Moves the Internet back towards its traditional model ("smart host,
   dumb network")

B) Router changes
+ The number of backbone routers is much smaller than the number of hosts.
+ Decision matrix is potentially much smaller (2 big core router vendors
   plus 3-8 big ISPs)
-  Feature can't deploy unless the big ISPs agree
-  We still have a lot of state in the core of the network; we are just
   handling and using that state a bit differently
-  Continues the current trend towards the new model ("dumb host,
   smart network")


Yours,

Ran Atkinson
rja@extremenetworks.com


--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg