Re: snmpconf advancement

I may be interested in collaborating with others on an open-source
implementation.

My game plan would be:
1. Start with the open source Javascript interpreter
2. Write the function library (this is quick and will demonstrate it is only a
small memory footprint)
3. Write the MIB code

At this point you have a running PM MIB mid-level manager/self-manager which is
probably too big to be embedded because of #1 which I presume would be large
(i.e. it isn't subsetted nor written with an embedded system in mind). But it
would be fine for a mid-level manager and for getting some experience. The next
two steps are:

A. Add a CLI function library
B. Write a PolicyScript interpreter from scratch to make it small. I think that
60K-80K is realistic.

The order of implementing A and B is subject to debate but I think that we'd
find it more interesting to do A first because this is just a toy without
access to controllable objects and SNMP doesn't hold a candle to CLI yet.

Anyone else interested?


Steve



Harrie Hazewinkel wrote:

> --On Wednesday, June 19, 2002 11:24 AM -0400 "Harrington, David"
> <[email protected]> wrote:
> >
> > I would also like to personally encourage open source developers to
> > produce a reference implementation, probably as a mid-level manager that
> > translates snmpconf scripts into mib-specific SNMP commands, so operators
> > can determine whether the language is sufficiently easy to use, and the
> > CPU overhead can be measured.
>
> Speaking as someone who works a lot with open-source. An open-source
> implementation gets done by two groups;
>    1) poeple doing it as study or hobby and
>    2) poeple who can do it because there employer supports it.
>
> So is Enterasys willing to sponsor such an effort?? Or others??
>
> cheers,
> Harrie