Plateforme de Calcul pour les Sciences du Vivant Summary of Embrace WP3 meeting Nicolas Jacq (CNRS) on behalf of Vincent Breton

Plateforme de Calcul pour les Sciences du Vivant

http://clrpcsv.in2p3.fr

Summary of Embrace WP3 meetingNicolas Jacq (CNRS) on behalf of Vincent Breton


Reminder: WP3 goals and structure

• WP3 goal: initiate a grid-based infrastructure for bioscientists– Making available most of the popular databases and software

products– Taking away the need for maintaining local copies of databases

and software– Taking advantage of the existing european national and

international grid infrastructures and the services they provide

• WP3 structured in 3 sub work packages– WP3.1 Technology recommendation and watch – WP3.2 test problems analysis, deployment and evaluation– WP3.3 technology adoption and application deployment


WP3.1 Technology recommendation and watch

• Objectives– Recommendation for technological choice to the Content

Integration work packages and to the Tools Integration work package at project month 18

– Technology watch during the project lifetime as grid technology is in constant and quick evolution

• Description of work– Evaluation of existing technologies for data and tools integration

and deployment

– Evaluation of existing technologies against user/provider

requirements


WP3.2 test problems analysis, deployment and evaluation

• Objectives: user driven technology recommandation– Test problems will be deployed before WP4 test cases – They will be used to evaluate existing technologies

• List of test cases– Distribution and update of biological data bases in a grid

environment, led by BMC– grid portal BioASP, led by CMBI – Computer intensive applications, led by SIB

• Description of work– Analysis of test cases – Deployment of test cases – Methodologies for evaluation – Evaluation


WP3.3 technology adoption and application deployment

• Objectives– To stand at the interface between the project and the grid

infrastructures. – To evaluate existing infrastructures for the deployment of

EMBRACE gridifed tools and data bases– Provide support to the technology adopted within the project

• Description of work– Evaluation of Grid infrastructures for bio-informatics applications – Security – Support to creation of virtual organisations and/or virtual services

– Support to technology adoption


WP3 roadmap

Due Date Number Title Partners involved

PM12 (1/2/06) D3.2.1 Initial technological analysis of test problems

CNRS, SIB, CMBI, Uppsala

PM18 D3.1.1 & M3.1.1 & D3.3.1

report on technology & network infrastructure survey & technology recommendation

All WP3 partners

PM24 M3.2.1 First implementation of test cases


PM24 D3.2.2 Second technological analysis of test problems


PM36 D3.2.3 First evaluation of test problem implementations



Meeting agenda

• Evaluation of underlying middlewares deployed on grid infrastructures in Europe– Identification of middlewares where WP3 partners have

expertise

• High level services and tools– Presentation of high level tools and services for end users

• Interfacing high level services and tools with underlying middlewares– Presentation of experiences with high level services and tools as

well as attempts to interface them with underlying middlewares

• WP3 test cases– Presentation of the test cases

• Open discussion


Session on underlying middleware evaluation

• Goal: evaluation of underlying grid middlewares that will be used to deploy data & integration tools on grid infrastructures

• Encourage WP3 partners (particularly test cases) to deploy applications on their local/national infrastructure and report on their experience– From the list of partners, the main middlewares should be tested:

LPC: EGEE( LCG2,gLite), SRB SIB: Swiss BioGrid (GT2 & GT4) Uppsala: Swegrid (Globus) Cineca: national Grid ( tbd) CSC: Nordugrid, DEISA CMBI: based on Globus, delayed CNB: EGEE EBI: Globus (GT4) / LSF

• Need for common evaluation criteria


Session on high level services and tools

• Integration tools– Presentation of Soaplab (M. Senger, EBI)

SOAP-based Analysis Web Services Further developments in Embrace WP2 in collaboration with Martin

– Presentation of BioMart (A. Kasprzyk, EBI) A join project (EBI, Cold Spring Harbor Laboratory) to develop a simple and scalable data

management system capable of integrating distributed data sources Further developemnts within BioMart and Embrace WP1

• High level services– Presentation of TAVERNA (T. Oinn, EBI)

Taverna is a workflow development and enactment environment Taverna acts as a client for several systems, Embrace being one of them Taverna expectation: Embrace is providing data sources Taverna tutorial for Embrace partners July 25th

• Are there other tools and high level services needed/requested ?– Biomoby

Interest to integrate biomoby standards specific interest for the BioMoby use cases which could be used as a model for Embrace use

cases – Hitkeeper, Blast


Session on Interfacing high level services and tools with underlying middlewares

• Interfacing Taverna with EGEE (T. Glatard, CNRS)

• User experience with Taverna and Soaplab (E. Rossi, CINECA)

• Discussion on issues related to asynchronous web services (E. Deandres, CNB)

• Issues raised– Some present limitations of Taverna for multithreading and

pipelining will be addressed by moving to 2.0 with new workflow core by end 2005

– Dedicated component between Taverna and EGEE to optimize job submission and handle fault-tolerance


WP3 test problems

• Test problem SIB– 4 applications requiring heavy CPU – Access to more computing power to increase the size of the database– Issue: to optimize job partition and have it submitted to the best

infrastructure– To be demonstrated on swiss Biogrid / EGEE– Contact point: Marco Pagni (SIB)

• Test problem Uppsala: replication of databases– To be demonstrated on Swegrid – Use of specific ports for xNDT protocol– Contact point: Ann-Charlotte Berglund (Uppsala)

• Test problem CMBI: enable a developing bioinformatician in a small lab to make easily a simple in silico experiment and use existing servers in Europe– Have access to a registry of resources– Have libraries (perhaps for different languages) allowing to access

remote resources– Clients: students with basic knowledge (Python or Perl)– Deadline for deployment: end of third year


Test problems: issues

• The three test problems cover nicely service providers and end users needs on Embrace

• Important to get them deployed– Local deployment on european/national/regional grid– Evaluation of these infrastructures and their mw

• Important to express requirements and to list services needed for future deployment on Embrace Grid

• No need expressed to ask for deployment of Embrace test problems on existing infrastructures – SIB test case to be deployed on Swiss BIoGrid / EGEE– Uppsala test case to be deployed on SweGrid


Requirements database

• EGEE applications have generated a very large database of requirements– https://svn.lal.in2p3.fr/EGEE/PTF/web/requirements.html

• Possibility to contribute to this database as a project– Avoid reinventing the wheel and creating a tool– A lot of information already available– Contact point: CNRS

• Need to have people familiar with this database to identify relevant requirements for each test case– WP3 partners proposing test problems are encouraged to have

one expert to learn how to use the database– One Embrace representative to add new requirements to EGEE

database if needed


Open discussion / Issues

• Need to define a common language– Common understanding of

words like grids, middleware, services

• Need for an Embrace architecture– Hard to design an architecture

without use cases on which to validate it

– First attempt by T. Oinn (EBI)

• WP3 should promote Life Science Analysis Engine (OMG standard)

Credit: T. Oinn (EBI)


Conclusion

• Important meeting to set the scene and agree on the way forward– Priority given to the description and deployment of test problems

• Good representation of WP3 partners– CINECA (x1), CMBI(x1), CNB (x2), CNRS (x3), CSC (x2), EBI

(x6), SIB (x2), Uppsala (x3)

• Presentations available online soon from Embrace web site

• Next meeting in Clermont-Ferrand July 25-27th– Tutorial on Taverna and application deployment on EGEE


What WP3 can offer to WP4

• Now– tutorials (tools, application deployment on grids)– Help to access to grid infrastructures if needed

• Soon, help to formalize test cases– Templates & models

• At PM18, technology recommendation


What WP3 needs from WP4

• Description of test case– Absolutely necessary to have a formalized description of WP4 test

cases to use them in WP3– Recommended template: Biomoby use cases online at

www.biomoby.org• Requirements

– Prerequisite: a description of the test case according ot BIomoby template

– WP3 will help WP4 partners to express requirements

• Feedback on high level services and tools needed– Taverna– Soaplab– Hitkeeper, BLAST


Joint Embrace / EGEE Tutorial

• Where? Clermont-Ferrand (France)• When? 25th-27th July 2005• How to register? soon available from Embrace web site• Registration fees? NO• Preliminary agenda:

– Monday (Noon – 19:00) Taverna Tutorial

– Tuesday – Wednesday (noon) EGEE (Introduction to Grid, LCG2.x.x, gLite, Hands-on)

• Who? Embrace Post-docs, EGEE Biomed People• July 11th: 16 registered participants

Documents

Plateforme de Calcul pour les Sciences du Vivant Summary of Embrace WP3 meeting Nicolas Jacq (CNRS) on behalf of Vincent Breton