From: Chris Smith <address@hidden>
Reply-To: address@hidden
To: address@hidden
Subject: [Vrs-development] VRS Development Idea
Date: Fri, 22 Mar 2002 11:07:42 +0000
On Friday 22 March 2002 00:30, Bill Lance wrote:
> At this stage, we are still writing specs for the
> system. The Cluster Manager still has a lot of work
> to do on it. This includes refining the functions of
> the thing thru method api interfaces thru its
> necessary levels. But much of the general Cluster
> management strategy and basic logic also needs to be
> refined first.
I meant to speak to you yesterday about this Bill.
I'm doing some work for the Government here along the lines of a SOAP based
XML gateway. This has a spin off in terms of dotGNU because it'll need a
network server architecture soon.
Now, assuming that Goldwater is the middleware and given my obvious
experience with it :o) and that the CM is going to rely on GWs network
handling, I was going to propose that:
a) Bill moderates the VRS in general and leads the RM sub-project
b) I lead the CM sub-project
This way I can probably get a skeleton framework up that contains a dummy
cluster management suite and a dummy resource management suite (Apache with
mod_perl and some perl GW clients and a bunch of 'place-holder' GWService).
This can be done very quickly (I'm doing most of it for my client anyway!).
These dummies can then be expanded to contain the functionality required on
a
modular basis.
The point is that we can start with a dumb LDS very soon, that accepts SOAP
requests and returns resources (It'll probably do XML-RPC too TBH).
Initially the RM would just pull files of disc, and even invoke perl script
as a pretend webservice, from which results are returned.
Having got this far, I can start sorting out Goldwaters Domain Handling AND
start integrating the CM with that.
And at the same time someone can build a GWServer that integrates pnetlib
and
the dotGNU C# VM (Rhys's effort) so we can start serving requests for C#
webservices, and build the LDS/VRS functionality up in the background.
I'm quite excited by this approach as its very team based, with the
development effort progressing on several important fronts at the same
time,
gives something people can download and play with early on and provides a
vehicle where ideas can be prototyped and experimented on.
What do you think??
Chris
--
Chris Smith
Technical Architect - netFluid Technology Limited.
"Internet Technologies, Distributed Systems and Tuxedo Consultancy"
E: address@hidden W: http://www.nfluid.co.uk
_______________________________________________
Vrs-development mailing list
address@hidden
http://mail.freesoftware.fsf.org/mailman/listinfo/vrs-development