Dear All,
Please find attached the weekly GridPP Project Management
Board Meeting minutes. The latest minutes can be found each week in:
http://www.gridpp.ac.uk/pmb/minutes/this-week.txt
(as well as being listed with other minutes e.g. 030113.txt).
John
Minutes of the 70th Meeting of GridPP PMB UCL – 12th March 2003
===============================================================
Present – Dave Britton, Nick Brook (video, am only), Tony Cass (video), Pete Clarke, Tony
Doyle, John Gordon (minute taker), Steve Lloyd, Robin Middleton
1. PROJECT MAP
Discussion based on version 10 (not circulated), status date 1 March so will probably reflect
what we show next OC. This should be some time in May - 14-21st would be good - TD to
confirm NEW ACTION.
1.1 Issues
DB emphasised that the present project management issues must be resolved
(via completion or a change notice) by Mid-April, such that the Project
Plan is in place for the OC."
CERN
* 17 issues raised with TC. Mainly concerned with LCG milestones. We need to align our
milestones with LCGs.
* currency fluctuations have resulted in CERN losing £500K of the GridPP contribution .
This means that the freedom to fund LTA based at CERN is significantly reduced.
* Need to be more careful about EDG deliverables during last year of project.
Experiments
* Atlas 3.1.3 demo of GANGA is late
* 3.1.5 – The provenance of the statement 'Several 10's of millions of events produced in the
UK' was questioned.
* 3.2.2 demo prototype CLI (skipped, merged with 3.2.5)
* 3.2.4 Remote grid Mass Storage write capability not delivered by EDG yet
* 3.3.2 LHCb has not introduced monitoring using R-GMA
* 3.4.2 CMS – storing >10TB at RAL. Late due to lack of appropriate middleware from EDG.
* BaBar status of 3.5.7 unclear
* 3.6.5 CDF delayed – new date required
* 4.1.4 JCG to give evidence of BaBar rebate
* 4.1.5 Tier A hardware by end March – change form needed
* 4.2.2 The use by LHC did not reach the 80% target due to lack of demand from the
experiments
* 4.3.2 – done - use GridPP6 talks by Tier2s as evidence
* 4.5.3 – interested cat 2 sites - done
* 4.6.2 – 3TB in UK probably done by CMS and LHCb – to be checked ACTION??
Monitoring
? 7.2.3 DB demonstrated his impressive automatic processing of staff effort from quarterly
reports. Long discussion on staff reporting.
2. TRAVEL BUDGET
Current status: looks on target to meet estimate of £214K for the year but £28K refund from
EU will reduce this.
3. EXPERIMENTS
TD had concern that experiments will need to provide a lot of information for the proposal
both for application interfaces and data requirements for the infrastructure as well as separate
experiment software bids. NB will stress this at EB meeting next week (NEW ACTION)
Should have discussion on role of EB Chair – possible buy out?
4. CB FEEDBACK ON GRIDPP2 PLAN
Feedback from CB meeting included:-
Dissemination. This should be increased in GridfPP2 with a specific post defined and an
expert recruited.
Funds for Tier2 hardware.
Experiment data handling should be funded outside of LHC experiments software line to
include requirements of other experiments. Much discussion on what this meant. No
resolution. Will be continued during proposal discussion.
5. PROPOSAL
Discussion on each outline document presented. Two questions concerning all documents.
How should we present Egee? No conclusion.
How should we deal with history of GridPP? Have an explicit introduction describing
what we have done so far. ]
5.1 Tier-0
The proposal is mostly for staff but must mention hardware. Hardware costs for Tier-0 are
very big but they are only asking GridPP2 for a small fraction of this. Position of Tier1 at
CERN is unclear
Exit strategy implicit, workplan part of ramp-up to operation phase when there is no hole in
CERN funding.
Discussion at CB supported proposal to set proportion of GridPP2 funding going to CERN
12.5%. Must guard against this being seen as UK's sole contribution to LHC Computing.
5.2 Tier-1
The detailed justification of RAL may seem overkill but it is required to allow comparison
with other potential bids to the PPARC call.
Need to add linkage to Tier-2 and Tier-0
The Tier1 is a long-term project, needing a commitment beyond the end of GridPP2.
Middleware support is required in the UK even if no middleware development is done here.
This should be quantified in Tier-1 or production grid section even though the effort will be
included in the middleware line. The support need not be sited at Tier-1.
It was proposed that the section listing named RAL individuals be presented in terms of
defined roles with named people as evidence of the ability to perform the required task.
5.3 Tier-2
The detailed information is required to convince PPARC that they need to invest to access
substantial kit at these centres.
Board to oversee whole Tier1/2. How is this different from EB? Or a User Meeting? Should be
moved to management section for further consideration. This ties into the proposed
appointment of an Operations Manager for GridPP2 (see later).
5.4 Tier-N spreadsheet
To gather input on Tier2s in a standard manner. Asks for information that probably doesn't
exist (e.g. disk usage in TB*days) but this is still worth doing as it exposes the need for the
information so it can be required in future. A missing element is a Tier-2 MoU outline which
can be presented as the basis for interaction between the Tier-N centres and the PMB.
5.5 Tier-N spreadsheet notes
5.6 Middleware
Hope to include reference to Geoffrey Fox's document for Tony Hey.
Have asked relevant people for input.
Beware exposing list of stakeholders without understanding the relationships.
5.7 Experiment Grid Interfaces
How do we progress this? The EB meeting next week will inform this discussion.
How do we include other experiments? How does this relate to rolling grants as an exit
strategy?
5.8 Project Management
CB happy with similar structure to GridPP. PMB with effective size of 9 has worked well. CB
backed Dissemination post with funding. Should be intimately connected with PMB. Possible
operations post (discussed later) could also join. TB develops into Grid Deployment Board. EB
into User Board. GDB plans rollout and operation of UK grid, UB provides check that centres
deliver to user requirement. Both represented on PMB.
Build on the proven success of our current PM tools.
5.9 Operations – Travel
Straightforward. Propose that the present mechanisms are retained, with outturns by area.
Overall request based on historical use.
5.10 Hardware Requirements
Will be summarised as part of proposal – reality check against total estimates from Tier-1/-
2s.
5.11 Dissemination
5.12 Testbeds to Production
This section should describe a proposal to develop a production quality grid by the end of
GridPP2 through four interconnected lines:-
* Participating in LCG Grid
* Working with UK Core eScience Grid
* Interworking with grids of other experiments (eg BaBar)
* Participating in Egee.
Emphasis will be placed on LCG Grid. There is a case to define a specific post to manage the
move to a production grid during GridPP2.
PC replaces TD as second author of this section. JG to produce outline (NEW ACTION)
ACTIONS AS AT 12th March 2003 (previous actions not reviewed at this meeting)
=============================================================================
ACTION 64.1 : JG to review UK input to LCG.
ACTION 66.3 : TD to provide final document structure for GridPP2
proposal.
ACTION 66.4 : RM to talk to Keith Jeffery re- whether middleware
research and development is appropriate to the complex problem solving
areas of EU funding.
ACTION 69.2 : SL to contact the organisers at Durham to find out what
provision is available for a demo.
ACTION 69.7 : Tier 0: TC & TD to provide input document with JG & SL
reviewing.
ACTION 69.8 : Tier 1: JG & DB to provide input document with TC & SL
reviewing.
ACTION 69.9 : Tier 2: TD & SL to provide input document with JG & TC
reviewing.
ACTION 69.10 : Middleware: RM & PC to provide input document with JG &
NB reviewing.
ACTION 69.11 : Experimental Grid Interfaces: RB & NB to provide input
document with PC & DB reviewing.
ACTION 69.12 : Management: DB & TD to provide input document with SL &
RM reviewing.
ACTION 69.13 : Operations: RM & TD to do financial input documents
with SL & DB reviewing.
ACTION 69.14 : Hardware Requirements: NB & RB to provide input
document with JG & SL reviewing.
ACTION 69.15 : Dissemination: SL & TD to provide input document with
DB reviewing.
ACTION 69.16 : Testbeds Tier0, 1 & 2 - JG & PC to document input and
DB & TD to review.
ACTION 70.1 : TC to discuss Virtual Collaborative Environments with Les
Robertson.
ACTION 70.2 : TD to convey the board's response to AP regarding Virtual
Collaborative Environments.
ACTION 70.3 : TD to create a GridPP2 planning web page.
ACTION 71.1 TD to initiate proposal for date of next meeting with Oversight Committee
ACTION 71.2 NB to explain to experiments what they need to input to proposal.
ACTION 71. JG to produce outline for section on 'Testbeds to Production'
|
