The OMII-Europe Project and the Impact of European Union Funding

1
The OMII-Europe Project and the Impact of
European Union Funding

• Dr Alistair Dunlop
• University of Southampton

2
Talk Overview

• EU Funding
• Whats the Motivation
• Preparing a bid
• Running a project
• OMII-Europe
• How has EU funding shaped OMII-Europe?
• An overview of the project and its constituent
  parts

3
Motivation - Why pursue EU funding for
OMII-Europe? I

• Preface
• OMII-Soton already exists with goals to
• Reuse, reengineer and integrate grid software
  components that have already been developed
  within the UK e-science programme
• EU officials liked OMII model and wanted an
  equivalent within EU
• Also exists within USA in NMI
• Hence a call was created within FP6 (framework
  Programme 6) to fund an equivalent entity

4
Motivation - Why pursue EU funding for
OMII-Europe? II

• Primary reasons to lead a proposal
• Protect OMII brand
• We were expected to do it
• If we didnt do it someone else would have
• Addresses sustainability of institute through
  diversity of funding sources
• Primary reasons why NOT to LEAD a proposal
• Want money to do our own interesting research

5
Prerequisites before starting to preparing an EU
bid

• 1 Find the right EU call (cordis.europa.eu) and
  read the call, rules, requirements
• Requirements include specifying
• Closing Date
• Composition of Consortium (Number of partners and
  which member states)
• Instruments available (the format needed for the
  proposal)
• 2 Go to Brussels to talk to the relevant unit
  head/s to find explore a proposed call
• More subtle requirements
• Approximate budget available/suggested
• Which institutions to include/exclude
• Tone of proposal
• Level of competition

6
Before writing a proposal

• Meet with prospective partners to
• Understand what they want to do but dont
  agree!
• Get commitment from them to help and not compete
• Describe the level of budget available and their
  share
• Ensure they understand the EU rules as this
  impacts what partners can do
• In FP6 research activities are 50 funded. (11)
• In FP7 this is 75. (31)
• Sketch an outline proposal that conforms to the
  EU instrument, identifying partners to activities
  and partner funding

7
Outline proposal

• The outline proposal for OMII-Europe was not
  exactly as first thought
• The Instrument dictates certain tasks in the
  proposal
• I3 (Integrated Infrastructure Initiatives) muct
  include Network Activities, Service Activities
  and Joint Research Activities
• Politics dictates inclusion of certain partners
• The proposal has to be written to fit partners
  not the other way around
• gt You can shape a proposal but you cant dictate
  it

8
The full bid

• Preparing the Bid
• Admin part (takes a lot longer than you would
  expect)
• Part B or the Description of Work
• Approx 100 150 pages of text conforming to
  template describing what will be done
• All partners must sign this off
• Takes approximately 3 months of pretty much
  full-time work
• After closing date 2 months to hear if proposal
  is shortlisted
• gt proposal is fundable
• Hearings 6 weeks thereafter to rank fundable
  proposals
• Outcome 6 weeks later

9
Before you start

• Negotiations with commission
• Take at least 3 months
• Could cut budget and insist on changes to
  proposal (not technical)
• Need to complete a consortium agreement allow
  at least 3 months
• Total time from submission to start is approx 8
  months to 1 year

10
The good and bad news

• Good News
• Finance department is experienced and very
  helpful in managing finances of EU grants
• Plenty of other people who can help
• Gives you plenty of time to get a good project
  manager in place!
• After you lead one project you make a name for
  yourself and get invited to participate in others
• Good news for individual researchers employed on
  project travel, salary, etc...
• Bad News
• Large administrative overhead with little
  resource
• The admin costs are lumped with your research
  costs so to achieve balance you forfeit some
  research funding

11
Impact of EU funding on OMII-Europe

• Emphasis changed from Reuse, Reengineer and
  Integrate grid components to Interoperability
  and Quality Assurance of grid components due to
  partner contributions
• The partner list resulted in grid components and
  grid middleware distributions being included that
  were not initially considered
• The Instrument for submitting the proposal
  required the inclusion of tasks that were not
  foreseen

12
An Overview of the OMII-Europe project

• EU funded FP6 project (RI)
• Starting May 2006, initial 2 year duration
• 16 partners (8 European, 4 USA, 4 Chinese)
• Open Middleware Infrastructure Institute for
  Europe
• Complimentary to existing national programmes
  (OMII-UK, NMI, C-OMEGA, OMII-China)
• Goal is to provide key software components for
  building e-infrastructures
• Project will demonstrate proof of concept with
  expectation for a follow-on project in FP7

13
What will OMII-Europe do?

• Initial focus on providing common interfaces and
  integration of major Grid software
  infrastructures
• Common services
• Database Access, Virtual Organisation Management,
  Portal, Accounting, Job Submission and Job
  Monitoring
• These represent many of the outputs from the
  standards function groups
• Capability to add additional services
• Emphasis on porting and re-engineering work, not
  developing from scratch
• Infrastructure integration
• Initial EGEE/UNICORE/Globus interoperability
• Interoperable security framework

14
OMII-Europe guiding principles

• Committed to standards process
• Implementing agreed open standards and working
  withstandards process (GGF/Oasis)
• Quality Assurance
• Published methodology and compliance test
• All software components have public QA process
  and audit trail
• Working with similar projects and organisations
  to agree policies
• Impartiality
• OMII-Europe is honest broker providing
  impartial advice/information on e-infrastructures

15
What will OMII-Europe deliver?

• Repository of open-source, quality assured
  software services for gLite/EGEE, Globus,
  UNICORE and CROWNgrid
• Some services bundled with major grid
  distributions
• Initial integration work with EGEE, UNICORE and
  Globus
• Public reports on grid infrastructures
• Initial benchmark results
• Impartial advice and information
• Evaluation infrastructure to test services
• User support and training for services

16
Who benefits from OMII-Europe?

• E-infrastructure providers
• Choice of grid software to deploy can be
  determined by selecting the most appropriate
  system to manage resources.
• Achieved through common interfaces and
  interoperability of grid systems
• Decisions not constrained by membership of a
  particular VO
• Not required to deploy and manage multiple grid
  distributions
• E-science users
• Common methods for accessing grid Infrastructures
• Access to resources beyond the immediate
  e-infrastructure running a specific grid
  distribution
• Achieved through low level interoperability of
  Grid distributions
• Users not restricted to a specific, fixed set of
  resources
• E-science application developers
• Applications can be deployed and run on multiple
  grid environments through adherence to common
  services
• Not required to develop different solutions for
  different grids

17
Why Globus, UNICORE, gLite and CROWN?

• Minimal significant set
• gLite is a complete set of middleware developed
  within EGEE and is deployed to create a grid
  containing more than 150 sites and 30 countries
• UNICORE is a major EU and national middleware
  initiative and is deployed at many supercomputer
  sites, in particular those available through
  DEISA
• Globus is the world-leading open-source platform
  for Grid computing developed within the USA and
  is used for many research projects world-wide
• All three grid platforms have significant user
  bodies within Europe
• CROWNgrid is the middleware used on the major
  Chinese grid infrastructure

18
Database Service

• Implementation of the OGSA-DAI specification from
  the DAIS-WG within the Data function group of GGF
• OGSA-DAI service federates data resources with
  different support mechanisms (Relational/XML
  Databases/flat files) allowing uniform access
  across these resources
• Number of other data specifications emerging that
  may be considered later.
• transaction management byte IO Grid file
  systems etc
• DAIS implementation already available for Globus
  4
• Work is to port to UNICORE and gLite Alpha
  releases scheduled for May 2007.
• Evaluating OGSADai4UnicoreGS

19
Job Submit and Job Monitoring Service

• Implementation of the JSDL (job submission
  description language) and BES (Basic execution
  service) specifications from the Compute working
  group at GGF
• Common way to specify and control jobs
  (abstraction of O/S and cluster controller)
• Other specifications such as scheduling, but
  above are essential and well developed with
  implementations
• Work is to make BES and JSDL available on Globus,
  UNICORE and gLite
• Initial version for UNICORE available in May 2007
• JSDL translator (using XSLT) for gLite in testing

20
Virtual Organisation Management Service

• Authorisation service available for Globus and
  gLite.
• Provides information on the user's relationship
  with Virtual Organization groups, roles and
  capabilities
• Work to make VOMS available under UNICORE and to
  extend VOMS with SAML support
• SAML (Security Authorisation Markup Language)
  from OASIS Technical Committee. (standard for XML
  exchanging authentication and authorisation data
  between security domains)
• Alpha version for UNICORE with SAML support
  scheduled for May 2007

21
Accounting Service

• Implementation of the Resource Usage Service
  (RUS) from the management working group within
  GGF
• Tracks use of resources (accounting in
  traditional UNIX sense), but not concerned with
  payment
• Closely related to Usage Record (UG-WG) within
  GGF
• Specification available for public comment
• Alpha version of RUS (or equivalent) available in
  May 2007 for Globus, gLite and UNICORE

22
Portal Service

• Integration of the Gridsphere portal framework
  with Globus, UNICORE and gLite and provide
  portlets for job submit, accounting, etc
• Provide application level portability at a
  portlet level
• Portlets available for main OMII-Europe services

23
Additional Services

• Current solution
• Chinese partners will make all services available
  on Chinese CROWNgrid infrastructure
• In May 2007, launch of the second round of
  service integration

24
OMII-Europe JRA1 re-engineering activities
OGSA DAI BES VOMS RUS Grid Sphere Etc. Identified Components
EGEE (GLite)
UNICORE
Globus
Etc. OMII-UK, USA, China
25
OMII-Europe Infrastructure Integration

• This activity goes beyond the adoption of common
  services and focuses on full grid infrastructure
  integration through employing
• A common security infrastructure
• Much similarity (X509) and differences (handling
  of proxies, authorisation, anonymity and
  auditing)
• Intention to define a common security base
• Provide a strengthened form of X509 credential
  management through using Myproxy
• job migration between Globus/gLite/UNICORE
• Builds on Globus/UNICORE Grip project
• Close collaboration with GGF GIN WG

26
Project Structure and Effort Allocation

• Networking activities
• Management, Outreach, Training
• 8 Person Effort
• Service Activities
• Repository, QA, Support
• 25 Person Effort
• Joint Research Activities
• Re-engineering, new services, integration,
  benchmarking
• 67 Person Effort

27
Effort (Person Years) per Activity
28
OMII-Europe Project Partners

• 114 person years over 2 years, 5 million Euro, 4
  major Grid infrastructures

University of Southampton UK (coordinator) University of Chicago USA
Fujitsu Laboratories Europe UK NCSA, University of Illinois USA
Forschungszentrum Juelich Germany University of Southern California Los Angeles USA
Kungl Tekniska Högskolan Sweden University of Wisconsin-Madison USA
Istituto Nazionale di Fisica Nucleare Italy Beihang University China
Poznan Supercomputing Networking Center Poland China Institute of Computing Technology Beijing China
University of Edinburgh UK Computer Network Information Centre Beijing China
CERN, European Organisation for Nuclear Research Switzerland Tsinghua University China
29
OMII-Europe project summary

• Interoperability is difficult not because of
  technical issues but because it requires
  agreement
• No one wants to be seen to lose out to someone
  else
• OMII-Europe has support from the major grid
  Infrastructure providers to deliver
  interoperability
• No point to be trying to solve the problem
  without vendor support
• OMII-Europes emphasis on standards provides a
  non-biased approach towards interoperability
• An open independent process needs to be used to
  arrive at technical decisions
• Achieving interoperability is a long term goal,
  dont try and eat an elephant in one go!
• OMII-Europe will improve overall USABILITY of
  grid Infrastructures and improve INTEROPERABILITY
  of grid infrastructures over the next two years

30
Concluding Comments

• EU research budget is moving towards 2 of total
  EU GDP.
• Getting a share isnt that difficult but you need
  to be politically aware
• It provides a means to support real collaborative
  research within the EU
• Participating is far easier than being
  coordinating partner