From: "Warren G. Anderson" Date: Tue Feb 26, 2002 12:44:50 US/Pacific To: lazz@ligo.caltech.edu Cc: lsccompcomm@ligo.caltech.edu Subject: [Lsccompcomm] Minutes of LSC COMPUTING COMMITTEE meeting of 2002.02.26 Reply-To: lsccompcomm@ligo.caltech.edu Minutes of LSC Computing Committee Telecon, Feb 26, 2002, 11:00AM PST Minute recorded by W. Anderson ---------------------------------------------------------------------- Albert - found a toll-free number for these calls, 1-877-252-5250. Albert - wants input on spreadsheet - comments, etc Kent - suggests recognition of file services (as per his e-mail). Albert - asks people who are represented in the spreadsheet to send that info to add? Tom - let's add UTB for completeness. Albert - can Warren send a row to add to the spreadsheet? Warren - ok. Albert - wonders whether we should have a weekly updated site which shows resource usage a la McDonalds (e.g. 50 billion jobs served) at various installations. Alan - suggests someone develop this and then everyone else copies it. Kent - a GUI will be in the next release of LDAS to provide statistical info, but it will not be easily copied, it is encrypted. Suggests info could be supplied to a centralized location and posted. Albert - a longer topic - what do we want to do at the face-to-face meeting on Saturday? Tom suggested the topic of discussion for that meeting be "what is a tier II center?" Albert knows that others are worried about what the specific responsibilities and duties of Tier I and Tier II centers should be, and wonders if there should be a discussion of what roles each should play. He thinks that Tom has put together a template for how to consider these issues. Tom - has put a skeleton together; what is the totality of computing resources available within the LIGO community, what is availability of these resources to LIGO (are they shared or mandated in part toward other tasks?). His template considers hardware, software (eg. OS, compilers, etc), and management support (eg. how much manpower is available for upgrades, maintenance, etc). He will prepare a template for the meeting. Patrick - wants to raise another topic: where does the buck stop, us? Who makes the final call on decisions about computing resources? Tom - let's assume we do. Albert - thinks we report to Barry and Rai as the leaders of the laboratory and scientific collaboration. Sees us as an advisory body to the directorates of these bodies, and believes our recommendations will be taken very seriously. He thinks that there is currently no definition of how to coordinate our resources to take advantage of excellent opportunities, and believes that this group can help develop a policy for that. Tom - adds that if we don't solve the problem, who will. Patrick - was wondering if we would have the power to implement the solutions we find? He likes Albert's interpretation and is satisfied with that role. Albert - comments that the perpetual problem is we are idea rich and people poor. This won't easily be solved directly. However, we can enable research groups to participate in solving the problem through NSF funding by identifying needs and pushing for funding to satisfy them. Tom - notes that there will be a ramp up from where we are to where we want to go. We are at the beginning of a building process. Albert - compares the situation to when a similar group met to formulate the data analysis white paper. The NSF gave the LSC/LIGO a mandate to provide an overview to justify and coordinate research and funding. The white paper was generated and evolved through five iterations. He believes it's up to us to define what we produce to enhance the definition of objectives in a similar way. He notes that there could be overlap with the white paper, but that the white paper often identifies needs but does not supply solutions. Alan - one of our goals could be to rewrite the chapter where the overlap occurs. Albert - wonders if FNAL has a template for addressing these issues? Tom - No, this structure has never been attempted before, because we are the alpha implementation of distributed data analysis. He notes that this changes our perspective as to how to operate in ways not forseen in previous analyses of this problem, e.g. Kent was not originally seen as having to support Tier II installations. Albert - we should also try to figure out what people will want and need; what are the data analysis functions that the community will need. e.g. different levels of data, different analysis models, etc. Tom - thinks that would be very useful. Albert - suggests as topics to be considered "how do we go back for second pass over the data?", "how do we incorporate data from other interferometers?", etc. Patrick - approves of this approach. Albert - suggests we start by each putting together a list of what we forsee the analysis needs of the community being. Tom - points out that some of that is in the white paper. Albert - agrees and says that he will review that as starting point for the list he brings to the meeting. Tom - thinks our job is to map the white paper requirements (and our own lists) onto the grid environment. Tom - let's discuss some administrative matters for Saturday. What time should we start on Saturday? Albert - suggests 8AM. Asks if the lab cafeteria is open for lunch? Tom - yes. Albert - dinner at 7PM? Tom - whenever the group wants, will try to get reservations at a tapas restaurant, although he is not sure how easy it will be to discuss on a crowded Saturday night. Erik, Kent and Albert are staying on West side, so they will pick up their passes at west gate. Tom will drop off passes at Hilton for people staying there. Milwaukee drivers will pick it up at the west gate. Alan - should we meet at Hilton for breakfast? Tom - too much driving for people on west side of lab, he will check if the cafeteria is open for breakfast. Tom - Warren and Sam are staying at Hilton? Warren - is. Albert - thinks Sam is too. Albert - Barry indicated that he was interested in the meeting ... can we arrange DCS for him? Maybe a mini-closeout so people don't have to listen all day. Tom - do we need video? Albert - doesn't think so. He will set up DCS for 5PM Central. Tom - can have a speaker phone available. Albert - other business? Kent - what is weather like. Alan - snow flurries now. Tom - forecast looks a bit better for weekend. -- +-| HOME |------| Warren G. Anderson |-------------------| WORK |-+ | 1360 Bluewing Circle | The University of Texas at Brownsville | | Brownsville, TX, 78520 | 80 Fort Brown, Brownville TX, 78520 | | PHONE: (956) 350-0968 | PHONE: (956) 574-6746 | | CELL : (956) 499-6447 | FAX : (956) 574-6726 | ++-----------------------+---+-----------------------------------++ | EMAIL: wganderson@utb.edu | WWW: http://dirac.utb.edu/~warren | +---------------------------+-----------------------------------+ _______________________________________________ Lsccompcomm mailing list Lsccompcomm@ligo.caltech.edu http://mm.ligo.caltech.edu/mailman/listinfo/lsccompcomm