I met with Doyle Weishar (GITI) and Brian Kettler (ISX). The
purpose of the meeting was to discuss my report
on CoABS-ALP Technical Exchange Areas (TXAs), and further steps to
be taken regarding CoABS-ALP technology interchange, with the background
that:
Doyle said that their basic approach to technology exchange was to make the Grid and Grid services available to other programs, and to cooperate to a reasonable extent with other programs in helping them use those components. They have targeted an end of summer timeframe for making a beta release of the Grid and an initial set of Grid Services (probably including at least a trader/matchmaker and logging service) available. They are also preparing documentation on how to use the Grid and the released services. They are interested in developing services that are relevant to other programs (and hence in getting input on Grid and Grid service requirements from other programs), and in defining TIEs between programs. However, given current resources (both time and funding), the ability to participate in technology exchange activities was necessarily limited. Ideally, any such activity with another program involving significant effort would obtain funding explicitly designated for that activity.
In discussing the specific TXAs described in my report, and Marshall Brinn's specific areas of interest, Brian indicated that he would be interested in feedback on AAML, and on language requirements for describing component capabilities for use in matchmaking (in ALP, this would specifically refer to cluster and plugin capabilities). Mobility is also on the short list of capabilities to add to the Grid (specifically targeting version 2 of the Grid prototype in the Fall). Doyle and Brian agreed that ALP's abilities to support workflow-like business processes and wrapping of legacy systems were important areas to provide for in CoABS. We agreed that a coalition scenario involving an operation in which it was necessary to dynamically determine and connect various systems not known in advance might be a realistic one for an ALP-CoABS joint TIE (e.g., for logistics support, one where it was necessary to find and interface with local or specialized suppliers who might have web-accessible interfaces to their ordering or scheduling systems) .
Regarding Todd Carrico's suggestion for an ALP design analysis, we agreed that this was certainly something that would need to be done. Doyle said he felt that CoABS would be prepared to cooperate to a certain extent with an ALP effort to use CoABS technology (particularly the Grid and Grid Services), but that much of an effort at the level of detail Todd described would have to come from either the ALP program itself, or be separately funded. (We also noted that my TXA document actually contains the beginnings of such an analysis, in terms of identifying key places where specific capabilities such as matchmaking might be added to the ALP architecture, and some of the issues involved in doing so). Part of such an effort would involve work within ALP to identify requirements and priorities for specific technical additions from ALP's own perspective, which CoABS participants could not necessarily provide anyway. While a certain amount of work between the two programs can go forward on an informal basis (e.g., the way Marshall Brinn and Brian Kettler have taken time to work with me), an intensive effort to integrate ALP and CoABS activities beyond the end of the current OBJS effort would require a certain amount of change in current directions (probably for both programs), and hence an explicit decision by Jim Hendler and Todd Carrico to engage in specific activities.
In the meantime, we determined that some reasonable next steps that would serve to at least continue informal interactions would be: