Preliminary Project Plan for GridPP3

1
Preliminary Project Plan for GridPP3

• David Britton 15/May/06

2
Boundary Conditions
Timeframe GridPP2 Sep 07 to Mar 08
GridPP3 Apr 08 to Mar
11 Budget Line Unknown exactly.
Scale set by exploitation review input.
(both from GridPP input and
Experiments) Exploitation Review input from
GridPP FY07 7,343k
FY08-FY10 29,302k Total
36,643k
3
LHC Hardware Requirements
GridPP Exploitation Review input Took Global
Hardware requirements and multiplied by UK
authorship fraction. ALICE 1 ATLAS 10 CMS
5 LHCB 15
Problematic using Authors in the denominator
when not all Authors (globally) have an
associated Tier-1. Such an algorithm applied
globally would not result in sufficient hardware.
GridPP has asked the experiments for requirements
and their input (relative to global requirements)
is ALICE 1.3 ATLAS 13.7 CMS 10.5 LHCb
16.8
4
Proposed Hardware

• The proposal from the User Board is that that
  the hardware requirements in the GridPP3 proposal
  are
• Those defined by the LHC experiments
• plus those defined by BaBar (historically well
  understood)
• plus a 5 provision for Other experiments at
  the Tier-2s only.

5
Proposal

• We propose to use the UB input to define the
  Hardware request (and not include alternative
  scenarios).
• We will note that these hardware requirements are
  not very elastic. Strategic decisions on the UK
  obligations, roles, and priorities will need to
  be made if the Hardware is to be significantly
  reduced.
• (Internally, we should continue to discuss how to
  respond to lower funding scenarios).

6
Hardware Costs
Hardware costs are rather uncertain. We have
previously quantified this uncertainty as 10 per
year of extrapolation (10 in 2007, 20 in 2008,
etc). Translates to an uncertainty of about 3.8m
in the proposal.
(Actual numbers here will be updated these are
a few months old)
7
Tier-1 Hardware
(Work in progress numbers are still evolving!)
8
Running Costs
(Work in progress)
9
Running Costs

• Running costs traditionally charged indirectly
  (at institutes and CCLRC). Normally averaged over
  larger communities which tends to be to the
  advantage of particle physics.
• We hope this continues as long as possible.
• Exploitation review input contained 1.8m
  running costs split between Tier-1 and Tier2
  which is only 50 of the current estimate.
• Should we avoid explicitly include running costs
  in the GridPP3 proposal (on the basis that it is
  not known how these will be charged)? Instead,
  include a footnote pointing out the assumption
  that running costs are funded by other mechanisms
  (SLA, FEC).

10
Tier-2 Resources

• In GridPP2 we paid for staff in return for
  provision of hardware, which is not a sustainable
  model. Need a transition to a sustainable model
  that generates sufficient (but not excessive)
  hardware, which institutes will buy into.
• Such a model should
• Acknowledge that we are building a Grid (not a
  computer centre).
• That historically Tier2s have allowed us to
  lever resources/funding.
• That Tier2 are designed to provide different
  functions and different levels of service from
  the Tier1.
• Dual funding opportunities may continue for a
  while.
• Institutes may have strategic gain by
  continuing to be part of the "World's largest
  Grid"

11
Tier-2 Resources
A possible model - GridPP funds 15 FTE at the
Tier-2s (same as Tier-1). - Tier-2 Hardware
requirements are defined by the UB request. -
That GridPP pays the cost of purchasing hardware
to satisfy the following years requirements at
the current year price, divided by the nominal
hardware lifetime (4 years for disk 5 years for
CPU).
E.g. 2253 TB of Disk is required in 2008. In
January 2007, this would cost 1.0k/TB. With a
life-time of 4 years, the 1-year value is
2253/4 563k.
Note This does not necessarily reimburse the
full cost of the hardware because in subsequent
years, the money GridPP pays depreciates with the
falling cost of hardware, whereas the Tier2s who
actually made a purchase, have been locked into a
cost determined by the purchase date. However,
GridPP does pay cost up to 1-year before the
actual purchase date, and institutes which
already own resources can delay the spend further.
12
Tier-2 Resources

• Sanity Checks
• Can apply the model and compare cost of hardware
  at the Tier-1 and Tier-2 integrated over the
  lifetime of the project
• Total cost of ownership Can compare total cost
  of the Tier-2 facilities with the cost of
  placing the same hardware at the Tier-1 (estimate
  that doubling the Tier-1 hardware requires a 35
  increase in staff).

Tier-1
Tier-2 CPU (K/KSI2K-year) 0.071 0.043 DISK
(K/TB-year) 0.142 0.107 TAPE
(K/TB-year) 0.05
Including staff and hardware, the cost of the
Tier-2 facilities is 80 of cost of an enlarged
Tier-1.
Question Would institutes be prepared to
participate at this level?
13
Staff Effort
Currently using the GridPP input to the
exploitation review as the baseline (with the
addition of Dissemination Industrial Liaison) .
GridPP3
GridPP2
14
Staff Costs
15
Tier-1 Staff
(Exploitation review input)
The staff required will be 15 FTE to run and
operate the CPU, disk, tape, networking and core
services as well as provide Tier-1 operations,
deployment and experiments support managed in an
effective manner. Support will be during daytime
working hours (0830-1700 Monday to Friday) with
on call cover outside this period. CCLRC may
provide additional effort to underpin the
service. In order to provide staff present
on-site for 24x7 (weekend) cover a further 5 FTE
(2 FTE) would be needed. 9 FTE in GridPP1
13.5 FTE in GridPP2
Management Service
Disk Service
Tape Service
File System Service
CPU Service
Deployment Service
Experiment Support
Middleware Support
Core Services
Operations Service
Security Service
Network Service
Other Service
16
Tier-2 Staff
(Exploitation review input)
Currently GridPP provides 9.0 FTE of effort for
hardware support at the Tier-2s (London 2.5,
NorthGrid 4.5, ScotGrid 1.0 and SouthGrid 1.0).
This is acknowledged to be too low and operating
the Tier-2s is a significant drain on
rolling-grant funded System Managers and
Physicist Programmers. Large facilities require
at least one FTE per site, whereas smaller sites
need at least a half FTE. On the basis of
currently available hardware an allocation for
HEP computing would be 4 FTE to London (5 sites),
6 FTE to NorthGrid (4 Sites), 2 to ScotGrid (3
sites) and 3 to SouthGrid (5 sites) making a
total of 15 FTE.
17
Grid Support Staff
(Exploitation review input)
From the middleware support side, at least one
FTE is required for each of the following areas
security support storage systems workload
management networking underlying file transfer
and data management systems and information
systems and monitoring (where an additional FTE
of effort is anticipated ensuring that our main
contribution to EGEE is supported in the longer
term). It would be inappropriate to reduce to
this level of effort abruptly at precisely the
time that LHC is expected to start producing data
in 2007. Rather it is advised to phase the
reduction to this level over FY08 and FY09
thereby sustaining a necessary and appropriate
level of support at this critical time.
there will remain core Grid application
interfaces supporting the experiment applications
that will continue into the LHC running period.
These stand to some extent independent of the
experiment-specific programmes, although they
serve them. A total of 7 FTEs is required for
these common application interface support tasks.
It should be noted that the proposed effort in
this combined area is a significant reduction
from the current effort in these Grid
developments of more than 30FTEs.
18
Grid Operations
(Exploitation review input)
In order to operate and monitor the deployment of
such a Grid, a further 8 FTEs of effort is
needed, corresponding to the Production Manager,
4 Tier-2 Regional Coordinators and 3 members of
the Grid Operations Centre.
19
Future Management
(Last CB Steves slide)
GridPP2 GridPP2 Beyond GridPP Beyond GridPP
Project Leader Tony Doyle 0.7 Project Leader 0.7
Project Manager Dave Britton 0.9 Project Manager 0.9
CB/Tier-2 Chair Steve Lloyd 0.5 Deployment Supervisor 0.4
Deployment Board Chair Dave Kelsey 0.3 Deployment Supervisor 0.4
Applications Coordinator Roger Jones 0.5 Technical Supervisor 0.4
Middleware Coordinator Robin Middleton 0.5 Technical Supervisor 0.4
Total 3.4 2.5

• Project Leader appointed by CB search Committee
• Others by Project Leader?
• 2.5 ? 1.5 over time?
• What about CB itself?
• What about Dissemination?

20
Dissemination
4. The bid (s) should a) show how
developments build upon PPARCs existing
investment in e-Science and IT investment,
leverage investment by the e-science Core
programme and demonstrate close collaboration
with other science and industry and with key
international partners such as CERN. It is
expected that a plan for collaboration with
industry will be presented or justification if
such a plan is not appropriate.
For exploitation review it was assumed
dissemination was absorbed by PPARC. Unlikely at
this point! Presently we have effectively 1.5 FTE
working on dissemination alone (Sarah Pearce plus
events officer). Want to maintain a significant
dissemination activity (insurance policy) so
adding in industrial liaison suggests maintaining
the level at 1.5 FTE.
21
Full Proposal (work-in-progress)
Compares with exploitation review input of
36,643k which included 1,800k running costs
excluded above.
22
Proposed Balance (work in progress)