Longterm sustainability of eInfrastructures in LA: the EELA2 model

1
Long-term sustainability of e-Infrastructures in
LA the EELA-2 model

• Bernard Marechal (CETA-CIEMAT UFRJ - Spain
  Brazil)
• Philippe Gavillet (CETA-CIEMAT CERN)
• Roberto Barbera (University of Catania and INFN -
  Italy)
• CCICT 2009 Conference
• Kingston - Jamaica, 16.03.2009

2
SCIENCE e-SCIENCE
Science is what I can teach to my son
e-Science is what he can teach me
3
OUTLINE

• EELA and EELA-2 a short review
• The EELA-2 model

4
PROJECT OBJECTIVES

• EELA (E-infrastructure shared between Europe and
  Latin America) under FP6 www.eu-eela.org/first-ph
  ase.php
• Bridge consolidated e-Infrastructures in Europe
  and emerging ones in Latin America
• Create and operate a transcontinental Grid
  infrastructure
• Address cost-effectively regional problems
• Enhance Research in developing counties
• EELA-2 (E-science grid facility for Europe and
  Latin America) under FP7 www.eu-eela.eu
• Set up a high capacity, production quality,
  scalable Grid Facility
• Ensure round-the-clock, worldwide access to
  distributed computing and storage resources
• Support a wide spectrum of applications for both
  European and Latin American scientific communities

5
RESEARCH APPROACH

• EELA
• Establish scientific network
• Enlarge and train (hands-on workshops)
  communities
• Establish a pilot Grid supporting
  proof-of-concept applications
• Close collaboration with other EC funded projects
• Grid Cluster deployment, definition and
  enforcement of operational schemes and policies
• EELA-2
• Expand the current EELA infrastructure
• Provide the full set of Grid Services needed by
  all types of applications
• Collaborate with NRENs and create Regional
  Operation Centres
• Support actively the creation of National Grid
  Initiatives (NGI)

6
MAJOR OUTCOMES / RESULTS

• EELA
• Provide European researchers early access to a
  new, well-supported e-Infrastructure and enable
  them to speed up the processing of scientific
  data
• Strengthen scientific communities in Europe and
  Latin America
• Successful building of a reliable and almost
  production quality e-Infrastructure in Latin
  America
• Grid technology skills acquired to autonomously
  support the Latin American Grid on the long term
• Entrance and consolidation of Latin American
  communities in worldwide collaborations
• Amplification of the e-Infrastructures relevance,
  blazing the trail towards Latin American
  e-Science initiatives and/or NGIs

7
EELA-2 IN NUMBERS

• 16 JRU-type partners (6 from Europe, 9 from LA
  and 1 International Institution CLARA) in 14
  countries (5 from Europe, 9 from Latin America)
• 53 institutions, most of them clustered in 9
  Joint Research Units (JRU)
• EC support of about 2.1 MEuro ( 300 KEuro from
  CIEMAT)
• 6 Activities (Management, Dissemination and
  training Application support Infrastructures
  services Network resource provision Development
  of services for applications and infrastructure)
• 30 Resource Centres 3000 computing nodes 700
  TB of storage
• Foreseen growth over the project duration 20
  in computing and 15 in storage
• About 50 applications selected (Biomedical, High
  Energy Physics, Earth Sciences, Climate,
  e-Learning, e-Government, e-Industry)

8
EELA-2 COUNTRIES / RESOURCES
France Ireland Italy Portugal Spain
14 Countries 16 Partners (9 JRUs) 53 Members
Argentina Brazil Chile Colombia CLARA
(International) Cuba Ecuador Mexico Peru Venezue
la
9
GLOSSARY

• Key words
• DCI Distributed Computing Infrastructure HPC
  Grid
• Local
• RC Resource Centre
• Country
• NGI National Grid Initiative
• GridNGI National Grid Infrastructure, supervised
  by NGI
• GOC GridNGI Operation Centre
• NREN National Research Education Network
• NOC NREN - Network Operations Centre
• International
• LGI Latin American Grid Initiative
• GridLGI Latin American Grid Infrastructure,
  supervised by LGI
• GSC GridLGI Support Centre
• CLARA Cooperación Latino Americana de Redes
  Avanzadas
• RedCLARA Latin American regional backbone
  network, operated by CLARA
• NSC RedCLARA - Network Support Centre

10
Some concepts

• Environmental protection is the process of
  making sure current processes of interaction with
  the environment are pursued with the idea of
  keeping the environment as pristine as naturally
  possible.
• Economic development is made of three building
  blocks, i.e. information, integration, and
  participation.

• Social development also known as social change,
  refers to
• change in social structure the nature, the
  social institutions, the social behaviour or the
  social relations of a society, community of
  people, and so on
• any event or action that affects a group of
  individuals that have shared values or
  characteristics.

11
and analogies
Key analogy Grids are complex ecosystems of
services sold and bought by virtual
communities.

• Real world
• Social
• Environment
• Economic

• e-Infrastructures
• Virtual Organizations
• Resource Centres Service Provi-ders ? National
  Grid Initiatives
• Middleware services, Application support, and
  Training

12
Other analogies(en.wikipedia.org/wiki/Society
and links therein)
Social evolution
Grid evolution
13
Timeline of EU and US Grids
OSG
GRID 3
GriPhyN, iVDGL, PPDG
Inter-operation
WLCG
EGEE 1
EU DataGrid
EGEE 2
EGEE 3
LCG 1
LCG 2
Jürgen Knobloch/CERN Slide 13
14
Grid Ecosystem evolution in EU
The system becomes sustainable only if a social
change occurs. By analogy, this means a change
in the way we consider and support VOs
15
and in Latin America ?
16
The Latin American Grid Initiative
(LGI)(http//documents.eu-eela.org/record/1119/fi
les/)
17
LGI Management
18
A possible structure of a NGI in LA
A strong liaison with the NREN must be
established. Local conditions permitting, NGI and
NREN can merge into a single entity.
19
Strategy and Lines of Action

• Well-defined strategy needed to address all
  subjects and all stakeholders, in parallel.
• Mandatory interconnected lines of actions with
  well-defined actors and targeted people
• Define the conditions of a sustainable e-Science
  Infrastructure
• Collect the largest support from the User
  communities
• Get full support from Decision Makers (DM) at all
  levels
• DCI being international by essence, develop
  actions in common with DCI-related projects with
  the same objective of long-term support.

20
Conditions of a sustainable GridLGI

• Establish the conditions of a sustainable GridLGI
  by
• Adopting a suitable architecture model
• Anticipating its implementation and evolution
• Deriving the Manpower Effort to operate and
  support it
• Preparing a transition strategy from EELA-2 to
  GridLGI.
• Status
• Actions
• LGI Several contacts with CLARA
• Actors
• EELA-2 Management
• Targeted people
• EELA-2 Members, DMs, other DCI projects
• Backup documentation
• DSA1.3 The Long-Term Latin American Grid
  Initiative Model and Operation Cost Estimate
• WNA1.1 Strategy towards Long-Term
  sustainability

21
Interaction with User Communities (1)

• Interaction with User Communities to
• Progress in the dissemination of DCI need /
  usefulness
• Use extensively the EELA-2 Infrastructure
• Look for (more) Applications of interest and
  potential impact
• Increase the amount of people promoting Grid
  computing
• Get new Partners / Members joining EELA-2.
• Progress in the creation of NGIs
• 50 Members in EELA-2 but lt2-4gt Institutions /
  country.
• ? Need support from the whole DCI community when
  discussing with the Decision Makers of a given
  country
• Get a wide consensus on all beneficial aspects of
  Grid computing and support to sustainable Grid
  Infrastructures
• Get unanimous agreement on the LGI model

22
Interaction with User Communities (2)

• Status
• Actions
• JRU / NGI status
• In place in Spain, Ireland, France, Italy,
  Argentina, Chile, Peru, Venezuela, Colombia,
  Portugal, Brazil
• Ongoing creation in Mexico, Ecuador
• Missing Cuba
• New Partners Panama (JRU), Uruguay (Single
  partner).
• Initiative of a User survey to collect
  Statements of Interests
• Structured in 6 blocks of questions
• Tell us about yourself
• Your experience about DCI in your own Research
  work
• Your perception of Distributed Computing
  Infrastructures (DCI)
• Tell us your opinion about DCI impacts
• Your perception of the need and usefulness of
  National Grid Initiative (NGI)
• Your perception of the need and usefulness of
  International Grid Initiatives (IGI)
• Status Under discussion in EGI_DS
• This meeting
• Actors
• EELA-2, Belief, DCI projects

23
Interaction with Decision Makers (1)

• Interact with Decision Makers to
• Disseminate the use uselfuness of DCI
  technology at all levels of Decision making
  processes
• Eventually convince Decision Makers that
• e-Science is the natural evolution of todays
  research
• DCIs are useful but even more, indispensable in
  many fields
• DCIs impact goes beyond e-Science up to the
  social and industry sectors
• DCIs operation support, as proposed, is
  affordable
• Present the sustainability approach via the NGI /
  GridLGI architecture model and the operation
  costing
• Work out optimized implementation scenarii in
  local context.
• Road map with precise objectives and progressive
  funding

24
Interaction with Decision Makers (2)

• Status
• Actions
• Decision Makers Days already held
• KoM Trujillo (Spain), 20-23th Apr. 2008
• Monterrey (Mexico), 6-8th Oct. 2008
• Quito (Ecuador), 8th Dec. 2008
• Bogotá (Colombia), 24th Feb. 2009
• In depth discussions (EELA-2 presentation,
  e-Science status in the country, prospectives,..)
• Production of dissemination materials (Posters,
  Flyers,..)
• Initiative of customized fact sheets for Decision
  Makers Days
• Actors
• EELA-2 (NA1, NA2), GridTalk, DCI projects
• Targeted people
• DCI projects ? DMs

25
Collaboration with Grid-related Projects (1)

• Collaboration with Grid-related projects for
  common actions towards sustainability
• Coordination of Grid-related projects EGEE-III,
  EUAsiaGrid, EUMedGrid, EUChinaGrid, EUIndiaGrid,
  EELA-2, BalticGrid-2 and SEE-GRID-SCI for actions
  towards
• Top level EC, EGI_DS, LGI
• Ground level User communities
• Maintain permanent concertation through regular
  meetings and direct contacts
• Discussing initiatives of common interest.
• Launch, follow up agreed actions

26
Collaboration with Grid-related Projects (2)

• Status
• Actions
• Concertation meetings at EGEE08 and ICT08
  (Belief)
• Follow up via mail exchanges, ICT08 Web site
  comment form
• Outcomes
• Comments to EC on the Key themes for the
  2010-2011 programme, in particular the transition
  from the on-going projects to a proposed
  sustainable model for pan-European grid
  infrastructure
• Comments to EGI_DS on the proposed sustainable
  model for pan-European grid infrastructure,
  called EGI
• Initiative of a User survey to collect
  Statements of Interests
• Initiative of a Customized Fact sheets for
  Decision Makers Days
• Guide for NGI creation and metrics of progress
  assessment (SEE-GRID- SCI)
• Actors
• Collaborative projects, Belief, GridTalk
• Targeted people
• All stakeholders from ground to top level

27
Conclusions

• EELA-2 is striving to take all measures to
  prepare the long-term future of its
  Infrastructure
• In all aspects concerned
• With all relevant partners
• In collaboration with all similar projects.
• Sustainability require that some conditions MUST
  be fulfilled i.e.
• The usefulness of DCI (i.e. EELA-2) has to de
  demonstrated by the Users communities themselves
• A cost effective architecture model (NGI,
  LGI-based) is needed
• Existing JRUs should play a crucial role in NGI
  creation
• A realistic operation costing is needed
• The transition phase to a sustainable model is to
  be worked out
• User communities at large must provide full
  support to all initiatives, taking every occasion
  to discuss with local DMs
• With an open minded attitude a road map towards
  sustainability should be defined with the parts
  concerned (User Communities ? DMs).
• These conditions being met, the EC 2010-2011
  programme foresees SSA projects to help in the
  transition phase.
• The Caribbean could / should interact with us
  (EELA-2)

28
Operation Manpower Effort
29
LAST BUT NOT LEAST
RESEARCH IS NOT A COST..
RATHER AN INVESTMENT