27
1 Application status F.Carminati 11 December 2001

Application status

  • Upload
    dewey

  • View
    51

  • Download
    0

Embed Size (px)

DESCRIPTION

Application status. F.Carminati 11 December 2001. Application integration + Test plan 1 person / application. Integration assessment 1 person / application. Party!. Software integration. WP8. 8. 15. 29. 26. 3. 22. 5. 12. 19. 10. 17. WP9. Doc review WP1 M.Reale WP2 J.Templon - PowerPoint PPT Presentation

Citation preview

Page 1: Application status

1

Application status

F.Carminati 11 December 2001

Page 2: Application status

2Architect WP meetingNovember 10, 2001

Validation plan timeline – new

8 15 22 29 5 12 19 26 3 10 17

Doc review

WP1 M.Reale

WP2 J.Templon

WP3 I.Augustin

WP4 A.De Salvo

WP5 JJ.Blaising

Integration assessment1 person / application

Application integration + Test plan

1 person / application

TWG + WP8-10 General Meeting

Validation plans presented

WP8 staffVO set up

Basic testing (?)

Message to WP9 & WP10

Software integration

WP8

WP9

WP10

Party!

Start writing 8.2

Octob

er 3

0

2001

Page 3: Application status

3Architect WP meetingNovember 10, 2001

Validation plan timeline – now

8 15 22 29 5 12 19 26 3 10 17

Doc review

WP1 M.Reale

WP2 J.Templon

WP3 I.Augustin

WP4 A.De Salvo

WP5 JJ.Blaising

Integration assessment1 person / application

Application integration + Test plan

1 person / application

WP8 staffVO set up

Basic testing (?)

WP8

WP9

WP10

Start writing 8.2

Software integration

Party!

TWG + WP8-10 General Meeting

Validation plans presented

Page 4: Application status

4Architect WP meetingNovember 10, 2001

What next?The integration is now finished All the actors should be praised

However the integration process took 300% more time than planned May be this was physiological in such a large

project But it should not happen again We should not let it dis-integrate again

Single WPs must integrate regularly their new versions in the TB We need development and production environment They do not need to be released at once This has to be a continuous process

Page 5: Application status

5Architect WP meetingNovember 10, 2001

Application validation8.2 will contain all we can do from today to Christmas It is very important that now we have the

access as soon as possible This week one person/application Next week the validation groups

For the review we may have one more month of testing to report But WP6 must stand by supporting the users We may hint this in 8.2

We can still do a reasonable job, with a bit of luck

Page 6: Application status

6Architect WP meetingNovember 10, 2001

What we want

OS & Net services

Bag of Services (GLOBUS)

DataGRID middlewarePPDG, GriPhyn, DataGRID

HEPVO common application layer

Earth Obs. Biology

ALICE ATLAS CMS LHCb

Specific application layer

WP9 WP 10

GLOBUS team

DataGRID ATF

WP8-9-10 TWG

Mar

ch 9

2001

Page 7: Application status

7Architect WP meetingNovember 10, 2001

What we have

OS & Net services

Bag of Services (GLOBUS)

ALICE ATLAS CMS LHCb

Specific application layer

WP9 WP 10

GLOBUS team

DataGRID middleware

WP1 WP2 WP3 WP4 WP5

HEPVO common application layer

Earth Obs. Biology WP8-9-10 TWG

DataGRID middleware

WP1 WP2 WP3 WP4 WP5

If we manage to define

HEPVO common application layer

Earth Obs. Biology WP8-9-10 TWG

Common core use case

DataGRID middleware

WP1 WP2 WP3 WP4 WP5

Or even better

DataGRID middleware

WP1 WP2 WP3 WP4 WP5

HEPVO common application layer

Earth Obs. Biology WP8-9-10 TWG

Common core use case

It will be easier for them to arrive at

Page 8: Application status

8Architect WP meetingNovember 10, 2001

A modest proposal

Identify one / two experts from each applicationHave them meet regularly for some limited amount of time to produce a proposal Ideally a couple of months Meet in person or via videoconf

Discuss this proposal at the next architect – WP meeting Have the different applications accept this

proposal as their GRID baseline

Page 9: Application status

9Architect WP meetingNovember 10, 2001

Why this is fundamental?

The LCGP (Lhc Computing Grid Project) will require us to work on common projectsThe HICB (intergrid coordination board) expects proposal from the experiments It would be MUCH smarter to provide a single

core use case Instead of competing one with the other

The different GRID projects risk to diverge A common core use case could help them to

develop coherent solutions Or ideally complementary elements

Page 10: Application status

10Architect WP meetingNovember 10, 2001

Experiment activities

There is a quite large expertise in the experiments about GRIDExperiments are already using GRID tools in productionIt is important that this experience is put to work for DataGRID providing qualified feedback

Page 11: Application status

11Architect WP meetingNovember 10, 2001

CMS

Page 12: Application status

12Architect WP meetingNovember 10, 2001

CMS Production Sites, Data transfers

INFN

CERN

FNAL

Bristol/RALCaltech

Moscow

IN2P3

UFL

HIP

Wisconsin UCSDMin.Bias Objy/DB

.fz files

Objy/DB

RC archiving data

RC publishing dataDirect access to

INFN Objy Federations through AMS by V. Lefebure

GDMP widely used GDMP widely used

Condor-G used at few sitesCondor-G used at few sites

Page 13: Application status

13Architect WP meetingNovember 10, 2001

Job scripts – BOSS integration

“Produce 100000 events dataset mu_MB2mu_pt4” Request

decomposition(Job scripts)

JOBSRC

BOSSDB

Request monitoring(Job scripts)

ProductionDB

ProductionInterface

Production Manager

distributestasks to

Regional Centers

Farm storage

RequestSummary

file

RC farm

Regional Center

Data locationthrough

Production DB

Page 14: Application status

14Architect WP meetingNovember 10, 2001

CMS - MOP

Page 15: Application status

15Architect WP meetingNovember 10, 2001

Distributed DB TAG Analysis

Page 16: Application status

16Architect WP meetingNovember 10, 2001

ATLAS

Page 17: Application status

17Architect WP meetingNovember 10, 2001

USAtlas Tool Development

GRAPPA

Monitoring

Condor (G)

GRAM

GSI

MDS/GIIS/GRIS

GridFTP

Replica Cat

Replica Mgr PacMan

Packaging

Magda

Page 18: Application status

18Architect WP meetingNovember 10, 2001

Magda Architecture Diagramwww.usatlas.bnl.gov/magda/info

LocationLocation

Location

SiteLocation

LocationLocation

SiteLocation

LocationLocation

Site Host 2

LocationLocation

Cache

Disk Site

LocationLocation

Location

Mass

StoreSite

Source to cache

stagein

Source to dest

transfer

MySQLSynch via DB

Host 1

Replication task

Collection of logical

files to replicate

Spider

Spider

scp, gsiftp

Register replicas

Catalog updates

Page 19: Application status

19Architect WP meetingNovember 10, 2001

Grappalexus.physics.indiana.edu/~griphyn/grappa/

Page 20: Application status

20Architect WP meetingNovember 10, 2001

ATLAS run time environment & monitoring

atlasgrid.bu.edu/atlasgrid/atlas/atlas_cache/cache.html www.mcs.anl.gov/~jms/pg-monitoringheppc1.uta.edu/kaushik/computing/grid-status/index.html

Page 21: Application status

21Architect WP meetingNovember 10, 2001

LHCb

Page 22: Application status

22Architect WP meetingNovember 10, 2001

Globus use in LHCbGlobus-job-submit (tested, works in production) to: Testbed 0 Csflnx01.rl.ac.uk (RAL) Ccali.in2p3.fr (IN2P3)

Don’t use Globus RSL, give options on Globus-job-submit command lineSome instability in serviceNeed Globus client s/w on LXPLUSGlobus-rcp tested, not reliable enoughGlobus-FTP tests underway NIKHEF-SARA

Page 23: Application status

23Architect WP meetingNovember 10, 2001

ALICE

Page 24: Application status

24Architect WP meetingNovember 10, 2001

ALICE GRID August Production

Page 25: Application status

25Architect WP meetingNovember 10, 2001

ALICE GRID File Catalogue as a global file system on top of a RDBTAG Catalogue, as extension Secure Authentication Interface to Globus under development

Central Queue Manager ("pull" vs "push" model)Monitoring infrastructure

The CORE GRID functionalityhttp://alien.cern.ch

Page 26: Application status

26Architect WP meetingNovember 10, 2001

DataGrid & ROOT

Local

Remote

Selection

Parameters

Procedure

Proc.C

Proc.C

Proc.C

Proc.C

Proc.C

PROOF

CPU

CPU

CPU

CPU

CPU

CPU

TagDB

RDB

DB1

DB4

DB5

DB6

DB3

DB2

Bring the KB to the PB and not the PB to the KB

Page 27: Application status

27Architect WP meetingNovember 10, 2001

ConclusionTime is very tight for validation before the reviewThe release process will be of fundamental importance for the further development of the project

We have to follow it closely

Some work needs to be done to integrate and streamline products and procedures

But for this you need real users!

Some work needs to be done on the user side to provide a more usable/useful picture to developersA huge potential is there, up to us to exploit it correctly!