Agenda for the Call on Sep 27 Wed 20:00 EDT/Sep 28 Thu 9:00 JST
Folks, Here is an agenda for the first call after GGF18. See you there. ### Proposed Agenda 1. Role call, note taker & time keeper, agenda bashing. (5 min) 2. GGF18 wrap up and future (15 min) - Approval of the minutes - How to promote the ACS implementation. - Future ACS direction, extension and alignment with SDD. 3. Updates (5 min) If there are any others... 4. Wrap up, Next meeting(5 min) Total: 30 min ### Logistics for the call Date & Time: Sep 27 Wed 20:00 EDT/Sep 28 Thu 9:00 JST Dial-in Number: 1-712-432-2000 (Iowa) Participant Access Code: 7771111 -Keisuke
Hi team, Here is the minutes of the today's telecon noted by Mike. Thanks, Mike. Please review it and feed back corrections, if there are any. It is also available at below link: https://forge.gridforum.org/sf/go/doc13907?nav=1 -Keisuke ============================================================================== ACS Teleconference Minutes ------------------------------------------------------------------------------ Date and Time: Sep 27 Wed 20:00-21:00 EDT/Sep 28 Thu 9:00-10:00 JST Keisuke Fukui (Fujitsu) Michael Behrens (R2AD) - Notes ----------------------------------------------------------------------- GridWorld Slides: http://gridworld.pistonbroke.com GGF18 Slides: http://www.ggf.org/gf/event_schedule/materials.php?event_id=5 MyProxy and GT Slides: http://myproxy.ncsa.uiuc.edu/talks.html Regarding future directions, for AutoDetect: Network and assign role via one of these mechanisms: 1 - General Purpose 2 - Human in the loop 3 - Assigned by pre-registration - perhaps harder when large number of devices are in flux. In Human brain, hundreds of thousands of cells die every day. Could be an organizer or departmental server which is in charge of aligning the device. Multi-layer autonomy. Ebay notes that Manageability requirement was number one reason for purpose grid standards. Goal: Self-maintain data center Perhaps WSRT - or other standard could be used? Some things to consider in researching hardware/software: - Instant On PCs/Servers - Zen/VMWare - OpenBoot Keisuke believes that for the mega class data center with tens of thousand servers should be managed as self-maintained units of even smaller human cell like components - numbers of cells are dying and newly added every day without being noticed in a macro level; data center components should be treated like this and the services running should be insulated from the base infrastructure hardware. Auto Discovery and Auto Deployment (Provisioning) should be the parts of the technologies. Can detect something via network - MAC address - need to have an Orientation Agent. Might tie into scheduler? (1) AutoDiscovery (2) AutoDeployment for newly inserted blades - What existing type of technology exist today? - Combine this with ACS to complete the AP. Expand Use-Case in ACS Future. Need to take a look at Pete’s Use-Case. Next conference: Not determined. Group work on e-mail basis and the call will be schedule on demand.
participants (1)
-
Keisuke Fukui