On Sun, Jan 15, 2012 at 9:17 AM, Mattmann, Chris A (388J) <
[email protected]> wrote:

> Hi BW,
>
> On Jan 14, 2012, at 8:06 PM, B W wrote:
>
> >> No-where really -- it's just that we don't have an explicit
> representation
> >> of an
> >> architecture object in Apache OODT proper, as it existed in GLIDE (being
> >> built on top of PRISM-MW it afforded us the ability for architectural
> >> introspection, etc.)
> >>
> >> You mean the the tight coupling is a real characteristic but conceptual
> at
> > the messaging component that does not really exists as a concrete
> artifact?
>
> Yep, I think that's the key -- it's more conceptual in nature, but does
> exist
> in the form of whatever messaging layer substrate is chosen. At its heart
> Apache OODT is loosely coupled, IMHO, as it leverages the client/server
> architecture, which allows for e.g., dynamic additions of new clients, and
> servers; servers don't need to know which clients are connecting to them;
> clients need only the end points of the servers; clients and servers can
> go away, etc.
>
> Ok. Great! Thx.
But, no matter what concrete message layer substrate we choose the
synchronous nature imposed by OODT's architecture still exists.

Which I guess highlights the importance of advanced planning of reusable
scientific work flows that can be automated.

This is a mitigation tactic of the synchronous nature through staging and
checkpoints for fault tolerance and recovery?


Cheers,
> Chris
>
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> Chris Mattmann, Ph.D.
> Senior Computer Scientist
> NASA Jet Propulsion Laboratory Pasadena, CA 91109 USA
> Office: 171-266B, Mailstop: 171-246
> Email: [email protected]
> WWW:   http://sunset.usc.edu/~mattmann/
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> Adjunct Assistant Professor, Computer Science Department
> University of Southern California, Los Angeles, CA 90089 USA
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>
>

Reply via email to