Hi team, Here is the minutes of the last teleconf. Apologize for the late post. Please review it and point out where the record is not correct, if any. It is also available at Gridforge: https://forge.gridforum.org/projects/acs-wg/document/minutes-20050428/en/1 Don't miss that there is an action item for all! - Post your own issues list to the mailing list by the next call. Regards, Sachiko ====================================================================== ACS Teleconference ----------------------------------------------------------------------- Date and Time: April 28 20:00-22:00 EST 2005 / April 29 9:00-11:00 JST, 2005 Participants: Keisuke Fukui (Fujitsu) Peter Ziu (Northrop) Michael Behrens (R2AD) Motohiro Yamada (NAREGI) Hitohide Usami (NAREGI) Sachiko Wada (ASCADE) - minutes ------------------------------------------------------------------------ * Naregi updatges Brief explanation of NAREGI PSE architecture from Yamada-san. Q: Is the archive created manually or automatically. A: It is created automatically by the registration service. Q: Are there any use cases where output data are put back to ACS after execution? A: Not in the current architecture. But we may try to store output data in the future. However, even in that case, it's not sure to store the data in the ACS repository. Because we are now redesigning NAREGI PSE. We may adopt the pure ACS repository or simultaneous use of the ACS and the Application Pool that we already have in the NAREGI PSE. Q: How about the relationship with CDDLM? A: They focus on Grid/Web services while we focus on legacy applications and source code compilation. This point is a large difference. Q: Do you have multiple binaries for each of the target platform in a single archive?. A: No. They are registered as seperate archives and the Information Service manages the relationship between the binaries and the source code. Keisuke's comment: ACS will provide sort of key attributes so that users can identify each files. So it is possible to contain muitiple binaries in a single archive. The granularity of archives depends on users. * Date for F2F Date: May 23 and 24 Place: Washington D.C. - detailed information about location and lodging will be posted later. * Research updates on CDDLM, security, etc. + CDDLM Sachiko sent a summary of CDDLM specs and some questions to the CDDLM ml and have not received any response yet. More questions on CDDLM specs were raised: - How to deploy legacy application. Is a web services wrapper always required? - Can we deploy an application server using CDDLM services? - What is the relationship with WSDM and CDDLM? => CDDLM seems to use WSDM. + Security Mike looked into WS-I security profile and WS-Security. Agreed that we don't need to cover all of the security aspects in our spec since there already exist standards such as WS-I and WS-Security. The least things we should define are: - mechanism to sign the archives to detect modification - mechanism for authorization (e.g. how to define roles) * Issues list To avoid duplication of tracker items, collect items first, then select which items to be registered to the tracker. Action: post each own issues list with a short description to the mailing list by next call. * Next call Date: May 9 Mon 20:00 EST/May 10 Tue 9:00 JST Calling #: will be announced later