Re: [MGSN-Dev] MGSN system diagram

[James C. <jw...@st...>]

Andrei,

Thanks for the diagram.  It's a good reference point to work from.  Here
are some thoughts.

- Label the diagram with a revision number so we can refer to a specific
copy.
- It would be beneficial to have a separate purpose/functionality
description of the system as a whole, not just the components.  What is
the goal of this system?  This will boil down a set of specific
functionalities that we will define in GSML.
- We need to place this in the context of other work.
  What needs to be done with Mercury to enable this?
  What have others done?  Specifically, we need to summarize Andreas' and
  Bartosz's work.  Does your diagram overlap at all?  Any redundant
  systems?

Let's come up with a working system diagram, building off what you have,
along with a description of the MGSN.  Focus more on what the system does
rather than the components (since there are an infinite number of ways to
implement the functionality with the components).

It would be good to come up with some scenarios to show how it operates.
How does a user register?  How does a user schedule a station(s)?  How
does a station register itself?

So far, here at Stanford,  we have been working from the bottom up.
First, enabling low level control via GSML of station resources.  We have
session level control.  Next, the network.  So, what are the functions of
this network?  Once we know these (and we have some identified already) we
can begin to lay out an architecture and build it.

For example, from a bottom/up perspective, we have stations, now we need
to network them.  How do we find them?  First step, a registration service
that allows ground station operators (or their software agents) to
register their capabilities in the MGSN.  ONce we have that, we can begin
discussing scheduling and VGS capabilities.

--Jamie



On Thu, 6 May 2004, Andrei Aron wrote:

> I'm attaching the diagram in pdf and doc formats.
> I am in the process of creating a tentative attack plan as well.
> Jamie and Mike, please send me any input you have.
>
> Thanks,
> Andrei