Merit Network
Can't find what you're looking for? Search the Mail Archives.
  About Merit   Services   Network   Resources & Support   Network Research   News   Events   Home

Discussion Communities: Merit Network Email List Archives

North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

RE: CIDR Report

  • From: Roeland Meyer (E-mail)
  • Date: Wed May 17 13:44:11 2000

> Wolfgang Henke: Wednesday, May 17, 2000 10:16 AM

> This suggests a generally flatter architecture, maybe parallel.  Still
> remember the supercomputer discussions a few years back when experts
> argued if either KSR or Thinking Machines is going to be
> *the* big win?
>
> And today when you need a lot of processing e.g. for rendering, you
> simply stick a couple dozen Pentiums on a fast switched ethernet.

What you are suggesting here is some form of distributed routing
process. The current problem is that all the router protocols store all
their data in talbes and they ALL have near identicle copies of these
tables. BGP is simply a synchronization mechanism. In fact, all the
routing protocols can be viewed as simple synchronization mechanisms.
Re-architecting that structure is not going to be simple and could
probably be the basis of an IETF working group. If the IAB could ever
get off their political duff (and back to real work), maybe they could
lead such an effort.

These days, the IAB is too busy making political pronouncements that
have nothing to do with internet architecture. Inertia is great/huge and
that job is probably not do-able. Us mercenary commercial architects are
not going to do it for free either.






Discussion Communities


About Merit | Services | Network | Resources & Support | Network Research
News | Events | Contact | Site Map | Merit Network Home


Merit Network, Inc.