1 / 15

Defining France Grilles resource allocation strategy

Defining France Grilles resource allocation strategy. Gilles Mathieu, IN2P3 Computing Centre France Grilles International Advisory Committee – March 2011. What do we want?. We want to assess: How resources and services are distributed How needs are expressed and answered to

Download Presentation

Defining France Grilles resource allocation strategy

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Defining France Grilles resource allocation strategy Gilles Mathieu, IN2P3 Computing Centre France Grilles International Advisory Committee – March 2011

  2. What do we want? • We want to assess: • How resources and services are distributed • How needs are expressed and answered to • What we could do better • We want to have the means to: • Apply a national scientific policy • Better answer to user needs • Improve service delivery to the French community 2

  3. What can we do? • Define resource allocation principles • That consider user needs and national policies • That stay in line with Grid specificities • Apply these principles • Through a clear strategy at national level • Without interfering when not necessary

  4. Where do we start? • Get inspiration from… • What was done in EGEE • What is done in WLCG • How it works in other sectors (e.g. HPC) • How other NGIs do it • Build up a strategy that… • Makes the most of what already exists • Takes our specificities into account 4

  5. Proposed principles • Necessity to evaluate allocation demands • A priori – How much can we allocate? Does it follow national policy? • A posteriori – Have we allocated properly? Do we need to readjust? • Final decision… • Is taken by whom the resources belong to • Can result in formal agreements • Necessity to define the role of the NGI • As a facilitator between users and providers • As a possible point of contact and broker 

  6. Different use cases • Depending on who the user is • A new user or new community • An existing community for which we are the sole resource providers • An existing international community • Depending on what the user wants • Precise needs in terms of resources • Just "use the grid"

  7. Scientific committee Established VO NGI Operations NGI/VO Training Broker User RA to new users (1) • Proposed workflow Transfer request Matching community? yes Askresources no VO based resources allocation Is user "grid aware"? yes no Send user to Resources > threshold? yes Ask for validation no Transfer request Request agreed? yes Transfer request no Project redirection or rejection NGI based resources allocation

  8. NGI Operations Resource Providers Scientific committee Service Providers recommandations Site based support agreements Site based support agreements Site based support agreements Site based support agreements Services proposals Resources proposals negotiations RA to new users (2) • NGI based agreement Support user in: New VO User Resourcesrequest "Catch all" VO Needs for Resources Needs for Services agreement User

  9. RA to new users (3) • A posteriori analysis • What has been used? • Does this answer user’s needs? • Have agreements been followed? • Is there a need to renegotiate?

  10. RA to established communities (1) • With an existing allocation process • E.g. WLCG • We should not interfere or conflict • Without a formalized process • E.g. biomed • We should be here to help

  11. EGI VRC Strategic policies Facilitation • Concrete agreements NGI VO Sites RA to established communities (2) • Place the NGI as a facilitator for resource negociations

  12. Can we implement this? • “New user” use case • Establishment of a national VO under study • Modalities of establishing a NGI point of contact • “Established communities” use case • Study usage of a resource negotiation tool • Bring the discussion at EGI level

  13. What’snext? (1) • “new user” use-case through national VO • Validate initial hypotheses • Study pertinence and feasibility • Can we technically do what we want? • How does that currently work (e.g. local VOs)? • Reach an agreement with all involved parties • Establish national VO specifications • Define deployment plan • Timeline: Now – Fall 2011

  14. What’snext? (2) • RA to established communities • Validate initial hypotheses • Study pertinence and feasibility • Is something better than nothing? • Do we have the means to enforce our strategy? • Initiate a wider discussion (e.g. within EGI) • Reach an agreement with all involved parties • Define and organize definition tasks

  15. Open questions • How can we easily integrate our strategy to what exists? • Is there always a need for filtering? • How not to mix “allocation” and “reservation”? • Is it worth the effort?

More Related