RE: Draft of updated WG charter

[Jay Ford <jay-ford@uiowa.edu>]

On Fri, 9 Jan 2004, Ayyasamy, Senthilkumar  (UMKC-Student) wrote:
> >I'll go further out on that limb & say that any multi-homing solution which
> >requires substantially more intelligence in end systems than is currently
> >required for multi-homed IPv4 is bound to fail, even if you somehow get it
> >deployed.
>
> disagree. For example, the way I use multi-homing from school is to have a
> f5 system which load balance between the main connection (morenet, Internet2)
> and my wireless connection. I also know that the whole LCS building at MIT
> has a web proxy system which allows to do a kind of Load sharing between
> MIT and a 100Mbps cogent connection. Given this is a *site* multi-homing
> WG, it is a known fact that many enterprise networks use load balancing
> products for connecting to multiple providers. I can even argue that
> end-site can achieve a equivalent reliability by load sharing among multiple
> inexpensive broadband connections in addition to a single dedicated line.
> In short, we are moving from a phase where availability is increased not
> only by network level redundancy but also end site level redundancy.

Fine.  Each site has a single prefix.  The links & routing establish the
topology, even with load balancers, traffic directors....  No problem.

> >  o  as evidenced by worm attacks..., the end systems are the worst managed
> >      pieces in the whole puzzle run by users who don't (& I'd say shouldn't
> >      be expected to) understand the workings of the network;  predicating
> >      routing-type functionality on that platform is asking for trouble
>
> come on... who is accountable for the multiple failures, accidental/stupid
> misconfigurations, black lists, unwanted rate limitations(have you read the
> recent IAB concern regarding port filtering.)?

Ultimately: the managers of the hosts, which are mostly end users who aren't
up to the job.  The filtering... was only put in place to protect the broken
systems because they couldn't deal with it themselves.  This is further
justification for not laying more networking responsibility on the hosts.

> >  o  there are at least 3 (4? 5?) orders of magnitude more end systems than
> >     there are routers, so embedding significant pieces of networking
> >     functionality in end systems greatly increases the likelihood of
> >     trouble & even more greatly decreases the chance of consistent
> >     operation over time
>
> on the other hand, who has the incentive to deploy? It is only the
> end-sites.

So, we have:
   hosts: large number; direct incentive, but no will or ability
   routers: small number; indirect incentive & real ability
I'll take the routers.

> In short, if we go by your suggestion, we can close this WG and standardize
> a simple BCP in idr WG (which will detail how IPv4 site-multihoming is
> applicable for IPv6.)

Cool.  That was easy.  :^)

> I think this WG is not taking that path.

Again, bummer.

________________________________________________________________________
Jay Ford, Network Engineering Group, Information Technology Services
University of Iowa, Iowa City, IA 52242
email: jay-ford@uiowa.edu, phone: 319-335-5555, fax: 319-335-2951