ietf-openproxy
[Top] [All Lists]

Re: Efficacy of rule specification, processing

2001-06-14 10:59:30

My opinion is that some of the local environment info is essential
and should be part of OPES.  Certainly the writer of the service
may feel that if the service is deployed without consideration of
the local environment, it won't work correctly, i.e., the benefit
will be nullified.

Where the local environment is used solely by the local
administrator to control resources and configuration, that's not part of the
OPES charter, but, if OPES does its work carefully, it should
be an obvious how to use the framework for exactly those
purposes.

Hilarie

Markus Hofmann <hofmann(_at_)bell-labs(_dot_)com> 06/12/01 07:25PM >>>

"Yang, Lily L" wrote:

Do we want to go that way? I would say no for now. Lets start simple.

Absolutely - "as simple as possible but not simpler". And that's
exactly what we're currently trying to figure out - how simple is
still acceptable.

To start, let's try to answer the simple question I posted previously:
Assume we've the same service available locally (i.e. on the OPES box)
as well as remotely (i.e. on a remote callout server). Now, do we want
to be able to do something like "if the load on the local system is
above a certain threshold, call the remote service, otherwise use the
local service"?

Just want to get a feeling whether people believe that this would be a
requirement that we've to deal with in the OPES framework, or whether
these things (here: load balancing) are completely separate. Notice -
we're NOT (yet) asking for solutions on how to solve this within OPES,
but just whether we need to solve this within OPES (or maybe that's
handle independently and separately).

Comments?

-Markus