COmanage Developers List

[heather flanagan <>]

For discussion on tomorrow's call...

---------- Forwarded message ----------
From: Benn Oshrin <>
Date: Wed, Jun 9, 2010 at 12:02 PM


Let me play the "I'm new" card again and state what is probably obvious to everyone else, but with the added complexity of making up new terminology that a least helps clarify things for me.

First, I'll use "COmanage" as a brand, much like "Apache" doesn't specifically refer to anything, but instead there is "The Apache Software Foundation" or "The Apache HTTP Server".  I'll also use "COmanage Project" as roughly interchangeable with comanage-dev.

Within that brand, we have something like the following:

 (1) COmanage VO Management System: Software (existing or new)
    developed and/or packaged by the COmanage Project that is
    basically a frontend and the provisioning glue.  There is
    an ongoing responsibility to maintain the software by the
    COmanage Project.

 (2) COmanage Application Domestication Services: A group, not
    necessarily 1-1 with the COmanage Project (and may in fact
    be a VO!) that domesticates or works with vendors to
    domesticate applications.  Ideally, this operates as a series
    of one time efforts, with ongoing responsibility for any changes
    made remaining with the appropriate project or vendor.

 (3) COmanage as a Service: A service offering targeted mostly to
    small and midsize VOs, consisting of an implementation of
    the VO Management System, various domesticated applications,
    and appropriate infrastructure (database servers, etc).

 (4) COmanage in a Box: A VM-based packaging offering basically the
    same services as COmanage as a Service, but designed to be run
    in-house rather than hosted.

 (5) COmanage Integration Consulting: Some combination of
    documentation and engagements to help sites deploy the VO
    Management System, integrate it with existing domesticated
    applications, and provide guidance for domesticating local
    applications.  There is no actual deliverable COmanage product
    here, though there could be a reference implementation,
    perhaps hosted with COmanage as a Service.

Some further thoughts:

 (6) The VO Management System is common to all deployment approaches
    (ie #3-5 above), so regardless of whether we should be focusing
    on the service model or the needs of the large VOs, defining the
    requirements and roadmap for it seems to be a pretty critical
    step.

 (7) The COmanage Project could have varying levels of engagement
    with the different deployment approaches.

 (8) It is not clear to me that COmanage in a Box should be an
    immediate priority.

 (9) I think there is a pretty straightforward architecture and cost
    model for COmanage as a Service, and drawing it out would be
    a good next step.

So perhaps this is just a very long way of saying I see a lot in common between the service model and the large VO integration model, and we can make some progress on both of them more or less in parallel.

Thoughts?

-Benn-