Attachment B. Harbison's Elaboration of CMA

Date: Sun, 14 Apr 1996
To: caeti@nosc.mil
From: harbison@Onramp.NET (Karan Harbison)
Subject: Architecture Efforts

.....

The CAETI MINIMAL ARCHITECTURE COMPLIANCE UPDATE

Kirstie took some time to expand my understanding of the CAETI Minimal Compliance Architecture.

* The MCA definition will remain constant over the life of the program. In other words, it is not evolving.

* The MCA was defined in Kirstie's Community Meeting II slides.

- It is at the comm protocol level. The options are listed on her slides. - It does not deal with data sharing beyond the concept of a message.

- A component must be able to "receive" a message but does not have to "understand" the message. A component must "receive" the 6-10 messages. That means that the pathways must be in place to accept a message, but it is not necessary to "understand" the message. - A component must "reply" that is has received a message. (There is some debate on how this fits with the "ears and nervous system" analogy). In some sense if there is a reply, then some action must be taken to reply. This implies that the nervous system has responded with an action - which requires an actuator in addition to the ear and the nervous system. - There is no explicit concept of integration beyond the message interchange.

* The ITDs will require additional compliance items. For example, a database for sharing data may be the first compliance step.

This page is URL http://www.computationallogic.com/software/caeti/architecture/model/cma-elab.html