https://www.na-mic.org/w/index.php?action=history&feed=atom&title=Engineering%3ATCON_04_20_2005 2026-04-22T18:49:58Z Revision history for this page on the wiki MediaWiki 1.33.0 https://www.na-mic.org/w/index.php?title=Engineering:TCON_04_20_2005&diff=3925&oldid=prev 2006-12-18T13:32:15Z

<p>Update from Wiki</p> <p><b>New page</b></p><div>== Agenda ==<br /> <br /> The dial-in number for this call (Wednesday April 20th, 9 am PDT [noon EDT]) was sent out in email.<br /> <br /> The agenda will cover some of the points raised in an email thread within the Engineering core related to grid infrastructure and the LONI pipeline:<br /> <br /> # Overview of goal and potential timelines<br /> # LONI pipeline overview<br /> # Grid architectural issues<br /> ## authentication<br /> ## exporting jobs (single job vs. entire pipeline)<br /> ## resource discovery<br /> ### compute resources<br /> ### application resources<br /> ## workflow optimization<br /> # Compute resources available for the programmer's retreat<br /> # Action items...<br /> <br /> == Minutes ==<br /> <br /> '''Attendees: Bill Lorensen, Steve Pieper, Mike Pan, Jeff Grethe, Jason Gerk, Simon Warfield'''<br /> <br /> * Discussions on security<br /> ** Jeff discussed BIRN's utilization of GSI and how the BIRN authentication architecture and services allow for single sign on across multiple resources<br /> ** Michael discussed the Pipeline's use of J-DRAMMA and how the Pipeline server connects to the SUN Grid Engine as the pipeline user and how client's could be made to connect to other remote resources directly<br /> <br /> * Bill suggested a 'staging grid' for compiling code and running continuous/nightly tests of algorithms on large sample data sets<br /> ** Requires the ability to stage newly compiled applications to compute resource without hands-on administration<br /> ** Three levels of applications under development were mentioned in the context of utilizing grid resources to aid the development process<br /> *** Development - application run locally<br /> *** Testing - nightly builds will require the dynamic staging of applications<br /> *** Production - application can be installed in a permanent location in standardized distribution<br /> <br /> * Next Call<br /> ** A follow up call next Wednesday at same time (April 27th, 9am PDT/noon EDT)<br /> ** Continue the discussion of job launching and authentication. More specifically, how would the Pipeline talk to grid resources - 1) does the client connect directly to the resource and how does the client get the module information; 2) does the server connect and how does authentication get passed through<br /> <br /> * Action Items<br /> ** Compute resources for Programmers retreat<br /> *** Jeff and Jason will confirm that BIRN compute resource will be available, also will<br /> <br /> check with Phil P. about possible additional fallback resources<br /> <br /> * ** *** Mike will check with LONI about availability of compute resources</div>

Andy