SEEGRID Review & SEEGRID2 KoM

2 May 2006, 08:00 → 4 May 2006, 18:00 Europe/Zurich

Attendees: VV, FE

Minutes Minutes SEEGRID_FinalReview_3May06_Istanbul.pdf

[1.03 Project highlights/summary] * 2 EGEE applications running Q: which EGEE applications? A: biomed and hep (cms, atlas, lhcb) * human network * sustainability via national NGIs * management infrastructure * meetings organised * deliverables produced * infrastructures size: - 30 sites (450 cpus) - 20 sites (150 cpus) without EGEE sites * regional applications: VIVE, SE4SEE and a lot more planed for SEE-GRID 2 * dissemination - contact with other projects: BALTICGrid, EELA, BELIEF... * BiH, Al: introduced a number of 3rd parties Q: Is there a NGI? A: Not yet, still a lot of political problems. * trainings * migration of sites to EGEE Q: Do you plan to move all sites to EGEE? A: Two grids overlap. Sites move to EGEE when they mature. Small/test sites stay in SEE-GRID to avoid strict EGEE rules. Q: Where there big problems with EGEE middleware? A: No big problems since we didn't use Glite, only LCG. Q: What is the relation to EGEE? Is SEE-GRID a small project waiting for EGEE or a small project making EGEE? A: We do not develop middleware but all problems detected were reported back to EGEE. Q: Is there a list of feedback to EGEE? A: Not at the moment but we can create such a list for SEE-GRID 2. Q: Where is EGEE going? Is it becoming a centralised organisation or a grouping of different grids? A: It is not clear yet. Q: Should there be 1 NGI per country or a big one for the whole region? A: Countries should have an NGI but organise in federation and appear together to get more influence. * all milestones met * project priority is research/development [1.04 WP1] * ensured that everything is: - on time - withing budget * maintained notebook Q: Why is Fabricio still listed? A: When he left CERN activity was already finished -> he did the work so his name stays. * created consortium agreements * work with other projects (Belief - indexing deliverables, eumedgrid, eela...) Q: Diligent was waiting for Glite, can you help them? A: SEE-GRID doesn't do middleware development but we contribute where possible (eumedgrid is starting to use the SEE-GRID site database). Q: Why is 0 hours accounted for contact with other projects? A: Senior persons that do this work can't be charged on the project. * more management needed in the first year * stick to LCG rather than Glite * 3 additional deliverables created: - site installation manual - application survey - sustainability plan * financing - reallocation of resources as some partners spent more than other - 42 more man-months than planed * recommendations from the first review taken into account R1 - infrastructure is good but not many users -> VIVE, SE4SEE applications, survey of possible new applications R2 - investigate sustainability -> creation of NGIs, D5.8 R3 - BiH status -> 3 new 3rd parties R4 - Al status -> 2 new 3rd parties R5 - grid applications more focused on the grid paradigm -> VIVE/SE4SEE R6 - involve industry -> Turkish data mining company -> Albanian national electric company R7 - training events -> number of events for admins, users and politicians organised R8 - collaborate with other projects -> mentioned before R9 - track other grid initiatives -> workshop with iceage (july, Budapest) R10 - benefits of SEE-GRID -> D5.8 [1.05 WP5] * web site - private area (docs) - public information * mailing lists * documentation naming scheme * promotional package - brochure - presentation - poster * newsletters on project progress * posters for specific events * training material is useful for other projects * other sources of funding: - Unesco HP - Sun * reached - admins - users - policy makers Q: How to measure the success of dissemination? Does the industry want to use the infrastructure? Is the government helping you? A: Serbia got financing for the cluster from the government. Unesco also donated some computers. We need to track users after the training. Industry can't use the infrastructure because the network policy forbids it. Q: Search function on the web doesn't find some deliverables? A: They are not all public. Q: Why are documents private? A: Financial documents are private, most other are public. Q: Diligent is indexing documents, will there be full text search? [1.06 WP3] * migration of middleware and applications + testing * using LCG * considerable CMS, Biomed usage * local apps: VIVE, SE4SEE Q: Is it possible to be a member of more than one VO? A: Yes, VOMS allows this. Q: Is it possible to submit job to more than one VO at once? A: No, LCG middleware doesn't allow this [1.07 WP4] * operational infrastructure * network infrastructure * site database created * monitoring tools * google maps with site status * helpdesk for admins/users * pgrade as portal Q: Is the portal different than the one in EGEE? Are they aware of this one? A: Yes, different groups in EGEE are testing it. Q: Is the SEEREN infrastructure good enough? A: Some partners are still having network problems or slow network. [1.08 Demo] * Gridice * GStat * SFT * MonALISA Q: On the personal level, how do you handle resource sharing? A: Using VOs. Site decides the % to allocate to each VO. Accounting shows the real usage. As we go into production this will become a bigger problem. * applications - VIVE - SE4SEE * usage statistics Q: What is the grid argument for applications? Why does it need the power of the grid? Choice for future applications should be: - demonstrate the grid potential - solves a problem for the society - promotes the grid paradigm not the apps A: VIVE and SE4SEE are proof of concept, new apps will be better. Q: Map applications to the features that demonstrate grid importance and then decide. [1.09 NGIs] * BiH - need to reactivate the network Q: Why not use the money from the current Internet connection to build the network? A: It is not enough and not sustainable without government support. Q: What is the time for creation of the network? A: Not clear, some laws are still missing. * Al - financial problems * Bg Q: Did you start with physics users and now trying to switch to other groups? A: Physicists started using it first, now others are also interested. * Ro - created NGI document Q: You have better NGI than some of the European states * Serbia - plan to have more users - than government will have to support it Q: Why is NRN separated from NGI? A: It is not yet clear what is better. Some countries have it together some separate. If there is a strong computer centre it usually becomes NGI otherwise NREN usually gets this role. Q: All countries made progress towards sustainability. If there was not SEE-GRID 2 would you survive and what would happen? A: Maybe. We still need to include more organisations. Things would not be integrated/coordinated. [1.10 SEE-GRID 2] * new directions - before: regional -> country level - now: country -> regional level * more sites * more apps * sustainability - NGIs - national strategy determined - establish CAs, ROCs * establish regional/national user communities - create developers guide for the grid * upgrade infrastructure - target stability/interoperability * strengthen the human network * beneficiary partner: one that has no national funding yet [Reviewer comments] * collaboration beyond the project * plan for stop of SEE-GRID * visibility on the national/European scene * apps are not the most important thing but they need to demonstrate the grid paradigm * list of criteria for selecting apps (cpu, disk, usefulness...) * successful apps = display window * external (industry) partners to help with the selection of applications that are useful to them * consider having regional level NGI for SEE to have more power to propose something * important role on the European scene - provocator/debater towards EGEE/establishment