Deliverable D2.2 Interoperability and interoperation guidelines

Transcription

1 Co-ordination & Harmonisation of Advanced e-infrastructures Research Infrastructures Support Action Grant Agreement n SEVENTH FRAMEWORK PROGRAMME Capacities specific program - Research Infrastructures FP7 INFRASTRUCTURES Project acronym: CHAIN Project full title: Co-ordination & Harmonisation of Advanced e-infrastructures Grant agreement no.: Start date of project: 1 December 2011 Duration: 24 months Deliverable D2.2 Interoperability and interoperation guidelines Version 1.8 Due on: 30/09/2011 Submitted on: 03/10/2011 Organisation name of lead beneficiary for this deliverable: INFN Dissemination Level: Public Abstract: This deliverable provides an analysis of the questionnaire data and existing knowledge and proposed recommendations for regional organisational and operational interactions and suggested developments.

2 Copyrights The CHAIN Consortium More details on the copyright holders can be found at the following URL: CHAIN ( Co-ordination & Harmonisation of Advanced e-infrastructures ) is a project co-funded by the European Union in the framework of the 7 th FP for Research and Technological Development, as part of the Capacities specific program Research Infrastructures FP7 INFRASTRUCTURES For more information on the project, its partners and contributors please see You are permitted to copy and distribute verbatim copies of this document containing this copyright notice, but modifying this document is not allowed. You are permitted to copy this document in whole or in part into other documents if you attach the following reference to the copied elements: Copyright The CHAIN Consortium. The information contained in this document represents the views of the CHAIN Consortium as of the date they are published. The CHAIN Consortium does not guarantee that any information contained herein is error-free, or up to date. THE CHAIN CONSORTIUM MAKES NO WARRANTIES, EXPRESS, IMPLIED, OR STATUTORY, BY PUBLISHING THIS DOCUMENT. Document log Issue Date Comment Author V0.1 07/07/2011 Table of Contents circulated for comments F. Ruggieri, R. Barbera, G. Andronico V0.2 25/07/2011 Feedback received and integrated O. Prnjat, F. Ruggieri V0.3 25/07/2011 TOC updated F. Ruggieri V0.4 29/08/2011 First contribution included N. Sinha V0.5 09/09/2011 Draft circulated for new contributions F. Ruggieri V0.6 13/09/2011 Modifications to the structure of the document and contributions V0.7 16/09/2011 Several contributions included V0.8 26/09/2011 New contributions, Executive Summary & Conclusions V0.9 27/09/2011 New contributions and corrections V1.0 28/09/2011 Consolidated version for further revision V1.1 29/09/2011 V1.2 30/09/2011 V1.3 30/09/2011 Corrections and new contributions, clean up of the text. New intermediate version for last minute changes and contributions Reviewed version with new text and corrections + reformatting. O. Prnjat T. Banda, B. Becker, G. Chen, P.S. Dhekne, F. Ruggieri E. Yen, M. Paganoni, F. Ruggieri P.S. Dhekne, E. Yen, S. Qian, M. Reale, O. Prnjat S. Al-Agtash, L. Nunez, S. Jalife, B. Becker, M. Ngwira, M. Paganoni M. Ngwira, F. Ruggieri, P.S. Dhekne, L. Matyska, K. Koumantaros, G. Andronico S. Qian, F. Ruggieri O. Prnjat, F. Ruggieri V1.4 30/09/2011 New revision and new contribution F. Ruggieri V1.5 30/09/2011 New revision and new contributions R. Barbera V1.6 30/09/2011 Final cleaning up and formatting F. Ruggieri V1.7 01/10/2011 Figure 8 Updated some further reformatting R. Barbera, F. Ruggieri V1.8 02/10/2011 Still some corrections and Fig. 4 with better resolution O. Prnjat, F. Ruggieri, R. Barbera, L. Nunez 2/50

3 Table of contents Table of contents Introduction Purpose Terminology-Glossary Executive summary The State of the Art High-level overview of the questionnaire results The CHAIN Knowledge Base The CHAIN Application Registry The Regional Grid Infrastructures The European role and actions Common Approaches and Uniform Solutions National Grid Initiatives and sustainability issues Interoperations with EGI and beyond Interoperability of Middleware Specific Regional solutions Africa (sub-sahara) Asia-Pacific Central Asia China India Latin America Mediterranean and Arabian Peninsula Technology-specific follow-up activities NGIs and sustainability issues Interoperations Operations Co-ordination (ROC management) Helpdesk Infrastructure and User Support Monitoring and Accounting Security Core services Middleware and service deployment Interoperability Standards and widely adopted solutions Beyond the demos The VRCs perspective Regional-Specific Planning Africa (sub-sahara) Asia Pacific Central Asia China India Latin America and Caribbean Mediterranean, Middle-East and Gulf (Arab Countries) Conclusions /50

4 1. Introduction The CHAIN project, started on the 1 st of December 2010, aims to coordinate and leverage the efforts made over the past 6 years to extend the European e-infrastructure (and particularly Grid) operational and organisational principles to a number of regions in the world. CHAIN uses these results with a vision of a harmonised and optimised interaction model for e- Infrastructure and specifically Grid interfaces between Europe and the rest of the world. The project is aiming to elaborate a strategy and define the instruments in order to ensure coordination and interoperation of the European Grid Infrastructure with those emerging in other regions of the world (Asia, Mediterranean, Latin America and Sub-Saharan Africa). The project has launched in the first months a survey via an on-line questionnaire targeted to national and regional Grid infrastructures (see Deliverable D2.1: State of the art questionnaire submitted to the EC in March 2011 and available on the project s web site 1 ) in order to update the information related to those initiatives. The analysis of the answers to the questionnaire have been presented in the deliverable D4.1 - Specificities of the various regional e-infrastructures 2 which provides insights into specificities of regional e-infrastructures (with the emphasis on Grid infrastructures), their commonalities and major differences Purpose The purpose of this deliverable is to provide a further interpretation of the analysis of the questionnaire data with the scope of outlining the proposed solutions for regional organisational and operational interactions Terminology-Glossary This subsection provides the definitions of terms, acronyms, and abbreviations required to properly interpret this document. AAI Africa & Arabia ROC CAS CHAIN C-DAC CLARA CNGrid CNIC DCI DoW EC EDGS Authorisation and Authentication Infrastructures Africa & Arabia Regional Operation Centre Chinese Academy of Science, China Co-ordination and Harmonisation of Advanced e-infrastructures Centre for Development of Advanced Computing, India Cooperación Latinoamericana de Redes Avanzadas China National Grid Computing and Network Information Center, China Distributed Computing Infrastructure Description of Work Annex I to the GA European Commission Equivalent Domestic Grid Structure /50

5 EGI EGI-InSPIRE EIRO EMI EPIKH ERA FP7 GA GOS HPC MoU NARSS NKN NREN NGI OSG RC ROC RPs SAGA VRC European Grid Infrastructure European Grid Initiative - Integrated Sustained Pan-European Infrastructure European International Research Organisation European Middleware Initiative Exchange Programme to advanced e-infrastructure Know-How European Research Area European Commission s Framework Programme Seven Grant Agreement Grid Operating System of CNGrid, China High Performance Computing Memorandum of Understanding National Authority for Remote Sensing And Space Sciences, Egypt National Knowledge Network of India National Research and Education Network National Grid Initiative Open Science GRID Resource Centre Regional Operation Centre Resource Infrastructure Providers Simple API for Grid Applications Virtual Research Communities 5/50

6 2. Executive summary The CHAIN project exploits the gathering of the experiences and best practices of regional Grid Infrastructure projects co-funded by the EC in the last six years. This knowledge has been updated and expanded in the work plan of CHAIN by means of an on-line survey aimed at regional and country-level representatives. This deliverable contains the information extracted from these different sources mediated by the experience of the contributors who are all involved in e-infrastructure activities in their countries and regions. The aim is to analyse, from one side, what has been done so far and, from the other side, define the existing and future directions in which these infrastructures will/should evolve with a focus on interoperability, organisational aspects, and long term sustainability. Current plans for the regions and specific countries are also taken into consideration wherever available. The main contribution of this document is concentrated in a series of synthetic recommendations that are listed quoting their time frame (short, medium, long term) and strategic relevance/priority (low, medium, high). These recommendations, provided in concise tabular format, represent the reference for future actions and initiatives not only related to the project lifetime, but also to yet-to-come initiatives and proposals that will endorse them. The deliverable structure first presents the state of the art as a high-level overview of questionnaire results and current way of systemising and preserving and presenting this current information, in chapter 3. Chapter 4 presents the questionnaire results from the point of view of both technical solutions and from the regional point of view. Then, in chapter 5, the recommendations are presented similarly per technology solution, while in chapter 6 the recommendations are presented per region. The core recommendations in this chapter provide the basis for further project developments, as well as potential work beyond project lifetime. Conclusions are drawn in chapter 7. 6/50

7 3. The State of the Art A large amount of the existing knowledge about Grid Infrastructures outside Europe has been produced by a series of regional grid projects (EELA 3, EUAsiaGrid 4, EUChinaGRID 5, EU- IndiaGrid 6, EUMEDGRID 7, and SEEGRID 8 ) that have been co-funded by the EC during the last six years. One of their shared objectives is to export the best practices of European activities in the field and facilitate the interoperations of the European e-infrastructure with existing or emerging ones in other regions of the world. The expected impact is twofold: (1) extend the European Research Area (ERA) and (2) address the requirements of Virtual Research Communities (VRC) that need to work in an intercontinental context. The documents describing the findings and the activities of the previously mentioned projects are available on their respective web sites High-level overview of the questionnaire results To complement and update the information that these projects have provided as input to CHAIN, the project has launched a survey using an on-line questionnaire. The analysis of the different aspects related to Regional Infrastructures and their operational and organisational structure are thus based both on the previous experience and the answers to the questionnaire. A detailed quantitative analysis of the answers to the questionnaire has been already presented in the Deliverable 4.1; Specificities of the various regional e-infrastructures and here we summarised some of the conclusions and provided a further view into the data focusing more on the interoperation and interoperability aspects. Synthetically, we can present the results in the following structured list: Regional coordinating bodies/organisations that are candidates for peer-peer agreements with Europe/EGI are as follows: o After a long process to foster the creation of a regional organisation in the Mediterranean and Middle-East, ASREN 9 recently appeared on the scene as a legal entity and it will need some time to organise and consolidate its role and representativeness. The geographical area of interest includes North Africa and the Arabian Peninsula. o Africa, being a very large continent with more than 50 countries, has currently three separated organisational groupings: North Africa with ASREN, sub- Sahara East and South with the UbuntuNet Alliance 10, and sub-sahara West with WACREN 11. UbuntuNet is a legal entity bringing under its umbrella several very advanced national initiatives such as the one in South Africa. o Asia-Pacific sees no regional organisation yet, although there is a strong role of the Academia Sinica in Taiwan that is supporting WLCG. o China and India have their own initiatives and organisational specificities and can not be easily homogenised with other countries in Asia. o CLARA 12 in Latin America has now some years of consolidated experience in Networks and seems to be more advanced in the process of becoming the www,wacren.net /50

8 reference organisation also for Grids. Some concerns are however related to the existence of two separated ROCs. o There is no coordinated initiative in Central Asia although, with the deployment of the CAREN 13 network, opportunities are opening up for Grid and DCI activities. o Sustainability is still an issue for all regions, although a common vision towards the regional organisation of e-infrastructures seems to emerge with different situations in Asia-Pacific, China and India that have a mix of strong national initiatives and a limited regional (inside Asia) coordination. The operational structure is generally based on Regional Operations Centres (ROCs), but these appear sometimes not to be legal entities that can commit on SLAs/OLAs and/or sign agreements for the long term support of their services. In particular: o At least two ROCs 14,15 are running in Latin America, with Brazil and Mexico having specific circumstances. o Africa & Arabia ROC 16 is at the moment the only initiative that covers the North and sub-saharan Africa extending its role also to the Arabian Peninsula. o Asia-Pacific ROC is well established and managed by Academia Sinica Grid Computing Centre in Taiwan. o In China the Computing and Network Information Center of the Chinese Academy of Science (CNIC-CAS) is the operation centre of CNGrid and CSTNET (the national Network of CAS). Thanks to CHAIN, China ROC 17 is being established at least to the resources where the glite middleware is deployed. o The Department of Information Technology (DIT), Government of India, has funded C-DAC (Centre for Development of Advanced Computing) to deploy and manage the nation-wide computational grid GARUDA. In addition regional WLCG operates under the coordination of Department of Atomic Energy. A limited number of middleware are deployed in the addressed regions. In the Mediterranean, Africa, Asia-Pacific and Latin America glite (now UMD) seems to be predominant. This could, in principle, simplify interoperation and interoperability with EGI. Globus-based customisations are used by GARUDA in India and ChinaGrid in China, while CNGrid uses GOS, its own middleware. In several countries desktop grid computing is present with different flavours. The above information processed by the CHAIN project needs to be organised, updated and presented on the web with a user friendly interface. The scope is to expose the current state of the art to e-infrastructure managers, technical people, policy-makers, stakeholders and general public. The description on how this is technically done is the topic of the following two subchapters. 13 caren.,dante.net /50

9 3.2. The CHAIN Knowledge Base In order to systematise the current know-how about e-infrastructure, CHAIN has developed in its first year a Knowledge Base 18 which, to our knowledge, is the largest in the world and gathers information about half of the existing countries. As shown in figure 1, the CHAIN Knowledge Base has tree views ; per country, per site and a tabular view. Figure 1 The CHAIN Knowledge Base In the country view, a world map can be navigated down to a single country for which the following information is shown with related web links (see figure 2): Regional Network; National Research and Education Network; National Grid Initiative; Certification Authority; Identity Federation; Regional Operation Centre(s); Grid site(s); Application(s) /50

10 Figure 2 Country view of the CHAIN Knowledge Base Figure 3 Site view of the CHAIN Knowledge Base 10/50

11 In the site, a 2D/3D Google map shows the sites belonging to the various infrastructures and, for each site, a web link is provided to the page of that site on the GOCDB it is registered to (see figure 3). In the table view, countries are listed in alphabetical order and, for each country, all the information available in the database are shown (see figure 4): Figure 4 Table view of the CHAIN Knowledge Base It is worth underlining that, wherever possible, the Knowledge Base of CHAIN is fed dynamically with information retrieved by other sources and this ensures that it is up to date The CHAIN Application Registry As you could see in the previous sub-section, the Knowledge Base includes information on the applications available in the various regional e-infrastructure. These are all gathered in the CHAIN Application Registry 19 (see figure 5) which has been developed as a tool to foster awareness and intercontinental collaboration. Clicking on a given applications triggers the opening of a tabbed frame (see figure 6) where more information is provided about the application and its developers /50

12 Figure 5 The CHAIN Application Registry Figure 6 Detailed view of an application 12/50

13 In the second year of the project, the Application Registry will include the RUN PAGE information to forward the user to the Science Gateway where that particular application can be run directly from inside the web browser. An example is already available for MrBayes, an application for phylogenetic analyses, that can be run inside the EUMEDGRID Science Gateway 20 (see figure 7). Figure 7 View of the MrBayes application inside the EUMEDGRID Science Gateway It is worth underlining that the CHAIN Application Registry does not include (and thus complements) the EGI Application Database applications.eumedgrid.eu 21 appdb.egi.eu 13/50

14 4. The Regional Grid Infrastructures This chapter analyses the commonalities and differences that suggest uniform/common or alternative solutions differentiated by region. An initial review of what has been the role of Europe and the actions taken so far introduces the main topics in the following section The European role and actions Europe has a long history of leading activities in the field of e-infrastructures in parallel to what has been and is being done in the US where the term used is rather Cyber-Infrastructures. In the field of R&E Networks the National organisations have promoted European projects aimed at interconnecting national infrastructures in a European-wide infrastructure (GÈANT) and created a European organisation to manage it (DANTE). Grid infrastructures have made a different evolution starting with European initiatives and then implementing the national level with different speeds. The process is however culminating with the creation of a European Grid Infrastructure (EGI) that aims to drive the evolution of Grid Infrastructures in Europe. The process took several years (since 2000) and was implemented in phases through European Commission s co-funded projects (EU-DataGrid, EGEE, EGI-DS and now EGI-InSPIRE and EMI). In 2004, SEE-GRID - the first regional project dedicated to the South-East region of Europe has started. The SEE-GRID infrastructure was built upon the EU Research and Education Network GÉANT and its SE European segment SEEREN, interoperable with the wide-scale product-level grid infrastructure of EGEE, thus contributing to efforts to establish a Pan-European Grid-enabled e-infrastructure. A series of initiatives aimed at extending the European experience and best practices to other regions outside Europe followed: EELA, EUMEDGRID, EUChinaGRID in 2006, EUIndiaGrid in 2007, EUAsiaGrid in 2008, the second phase of EUMEDGRID and EUIndiaGrid and CHAIN in In many regions the European approach to Grids has been exported starting from green field, and the activities included an intensive training and high-level technology transfer. The knowledge transfer has been done both towards technical staff and users with the support of the ICEAGE 22 and EPIKH 23 projects and finally adopting a general schema of a three week events divided in two parts. The first week dedicated to site managers delivers a program of technical presentations followed by a hands-on session where grid sites are installed from scratch both locally and in remote locations. Two other weeks are dedicated to users that want to port their applications on the Grid and are supported by tutors and experts in a dense program of application porting. At the end a one day workshop is organised to show the applications that have been ported during the event and make presentations on the importance of e-infrastructures A Marie Curie Action /50

15 4.2. Common Approaches and Uniform Solutions The experience made so far suggests that there are indeed several cases in which common approaches and uniform solutions have been adopted without seriously affecting the regional specificities. This is indeed the case for the regional infrastructures which have grown following the European model with no previous activities or planning. In these regions the national plans have also followed the European model which is NGI/EIRO based National Grid Initiatives and sustainability issues The creation of National Research and Education Networks has generally preceded any Grid activities in many countries and these organisations already have reached the status of legal entities and a consolidated experience in managing and operating infrastructures dedicated to Universities and Research Organisations. This makes them the first-choice candidates for fostering a Grid infrastructure in their countries. Moreover, NRENs are generally very sensitive to the implementation of software and applications that make use of the full potential of their networks and frequently require a push in the evolution of bandwidth and services. Thus, their interest in participating and leading e-infrastructure activities is very high and it is then not a surprise that many regional projects have seen the participation of NRENs and that, in many cases, the NGI that has been created de facto coincides with the NREN. This organisational approach indeed simplifies and speeds up the creation process and avoids conflicts of responsibilities that could happen where the process has been different. Of course, this is not the case in all countries where NGIs have so far been established. Operating a high-speed communication network requires, in principle, very similar organisational structures. However, several differences exist, the main one being that while the NREN can be regarded as a resource (network) provider, in the Grid the resources are generally distributed among the customers and the role of an NGI is essentially dealing with co-ordination and centralised operations and monitoring. At a pan-european level, the two are distinct and EGI is dealing only with Grid services, while on the country-level both approaches are represented although the largest ones (France, Germany, Italy, Spain, UK) have separated organisations for managing NREN and NGI. This different organisational approach, however, does not influence the European operations, while it could have an influence in the peer to peer relations among the EU Grid Infrastructure and those of the other regions. The CHAIN project is currently compiling the existing documentation on NGI guidelines that emerged from the previous projects such as EGI-DS, the SEE-GRID series of projects, the EELA series of projects, etc. It is also collecting existing regional sustainability guidelines. The draft of this documentation is available on the CHAIN project wiki 24. At a national level, considering the questionnaire results, the National Grid Initiative progress outside Europe is limited. China and India have established a certain form of National Grid Initiative - albeit without clear juridical status, and fundamentally based on a national funding programme originated by the Government. In China, the organisational model is a project consortium, with an order of magnitude of 45 organisations involved with similar situation in India. In the Mediterranean region, Algeria, Morocco, Tunisia and Jordan have a form of NGI which is embedded in the NREN structure as described above, thus demonstrating the usefulness of this approach /50

16 In Latin America, only Columbia reported the creation of a NGI, in the form of a task force, but the region is exploring ways to organise the services replicating CLARA network e- infrastructure. In Asia-Pacific, Taiwan has an established NGI in form of a consortium. In Africa, South Africa is on a good path towards formalising its nascent NGI. Beyond CHAIN project coverage, New Zealand is working towards NGI but without any legal status and concrete form yet. Finally, Iran is on a good path towards formalising its nascent NGI. Overall, the national data point to the fact that in smaller countries the establishment of NGIs related closely to NRENs is a suitable approach, while in larger ones it becomes more difficult to coherently articulate different initiatives and have a strong umbrella initiative. From the regional perspective, the questionnaire results point to solid organisation in Latin America and Asia-Pacific, where several Joint Research Units (JRUs) are established. Furthermore, Latin America has very concrete plans for embedding the Grid operations into CLARA. Finally, the establishment of ASREN is a very concrete development and its activities should include coordinating the regional DCI work and the existing NGIs Interoperations with EGI and beyond EGI has already documented 25 the required interfaces for service management, monitoring, accounting, authorization, etc. The path technically described in the EGI document fits with the needs of a resource site or a NGI willing to be part of EGI and addresses the related obligations. However, we have to consider that in several cases a regional infrastructure or even a NGI cannot easily accept to be seen as a part of EGI for political, economical and technical reasons. It is very hard to imagine that China and India (just to cite two examples) will follow this path as it is in the case of OSG and TeraGrid in the US. Exceptions, such as those already mentioned, could follow the OSG example and make some sort of SLA/OLA or a simple MoU with EGI if there is a need for close and quick interoperation. This of course implies that such scenarios should be dealt case by case as the technical and political issues cannot be foreseen in advance. Although this approach, due to the limited number of cases, sounds sensible from the practical point of view, it does not address the technical and political issues that will be faced. The proposed simplification, however, misses the opportunity to make some recommendations that could help in addressing the issue also from the technical point of view. Repeating the process each time with EGI on one side and different actors on the other side will be in the long term not efficient. Moreover, peer to peer agreements between EGI and X and Y will not automatically solve the problems between X and Y (the transitive property does not apply in this case). CHAIN s approach is thus to understand the issues, outline the possible alternatives and propose a roadmap that, seen from the different positions, is acceptable for all parties. Interoperation of the regional infrastructures with EGI and among themselves would ensure the actual provisioning of a set of resources to enable worldwide scale e-science, thus enabling to address the global challenges researchers are faced to in many scientific domains today. Beyond the motivations for pursuing interoperability, as a starting point for a roadmap towards it, this sections aims at describing how the different regions are providing basic required functionality for e-infrastructures, and underlying commonalities in the provisioning of core functions for the management of the e-infrastructures, their implementation by means of grid middleware-based services, and support to user and sites. 25 MS414 Integrating Resources into the EGI Production Infrastructure ( and the previous version MS407 ( 16/50

17 Operations Co-ordination (ROC management) As stated in the previous section, the operational structure is generally based on Regional Operations Centres (ROCs), but these appear sometimes not to be legal entities that can commit on SLAs/OLAs and/or sign agreements for the long term support of their services. At least two ROCs are running in Latin America, with Brazil and Mexico having specific circumstances. Africa & Arabia ROC is at the moment the only initiative that covers the North and sub-saharan Africa extending its role also to the Arabian Peninsula. Asia-Pacific ROC is well established and managed by Academia Sinica Grid Computing Centre in Taiwan. In China the Computing and Network Information Center of the Chinese Academy of Science (CNIC-CAS) is the operation centre of CNGrid and CSTNET (the national Network of CAS). Thanks to CHAIN, a China ROC is being established at least for those sites where the glite middleware is deployed. The Department of Information Technology (DIT), Government of India, has funded C-DAC (Centre for Development of Advanced Computing) to deploy and manage the nation-wide computational grid GARUDA. In addition regional WLCG operates under the coordination of Department of Atomic Energy. ROCs are typically staffed with the combination of project-based support and local cluster operators. Typically, they provide a team of shifting personnel, acting as infrastructure supervisors on duty and Ticket Process Managers for the Support System. There are no formal OLAs/SLAs enforced to sites apart from the commitment by site administrators to keep the site updated and running Helpdesk Infrastructure and User Support In some cases, the regional helpdesks are based on xgus/ggus ticketing system for providing support to users and site administrators: a dedicated xgus instance directly hosted at KIT (Karlsruhe) is the most common solution. Relevant support units are then created and maintained by the project/region Monitoring and Accounting The available monitoring solutions include NAGIOS, GOCDB instances for site registration and static data, Dashboards, GSTAT portals, and real-time monitors. Wiki sites appear to be a practical solution for collaboration of the operational staff. It is an advantage as those tools are compatible with EGI solutions Security The security solutions are mostly based on X.509 PKI certificates and the usage of the VOMS server, to provide the AAI infrastructure. However, due to the inherent difficulty in using the middleware by non-expert users, especially the management of personal certificates, and to the need of simplified, yet secure, authentication and authorisation mechanisms, some very interesting attempts have been done at INFN Catania to create Science Gateways supporting Identity Federations based on the 17/50

18 SAML protocol and the Shibboleth implementation. According to the results of the REFEDS 26 task force of TERENA 27, 16 million users in the world have an e-identity registered in an Identity Federation and this could enormously increase the number of users of e- Infrastructures. It is worth noting, also, that Identity Federations are usually operated by NRENs and this could bridge even stronger network and grid operators. A map of already established Identity Federations is shown in figure 8. Figure 8 Map of existing Identity Federations The lack of Identity Federations in many areas addressed by CHAIN can/will be addressed in the second year of CHAIN and eventually in future projects/initiatives Core services Typical Core Grid services are provided by glite 3.2 based services: the CREAM CE, lcg-ce, DPM-SRM Storage Elements, glite based Worker Nodes, LFC File Catalogues, the WMS/WMproxy server for Job Management, MyProxy and the AMGA metadata catalogue. The regions based on other types of middleware have a different set of services. More details are provided in the interoperability section Middleware and service deployment The different middleware deployed in the various regions are discussed in the following section. 26 refeds.org /50

19 Interoperability of Middleware The already mentioned projects, co-funded by the European Commission, disseminated the glite middleware in the addressed regions. The glite middleware was widely adopted in many cases and especially where a local middleware was not present. In the remaining cases interoperability solutions were studied and, to some extent, tested. In Table 1 a list of the middleware and infrastructures reported in the CHAIN survey is shown. Table 1 Middlewares and infrastructures addressed in the CHAIN project Middleware Infrastructure Regions UMD EGI Europe, Latin America, Africa, South East Asia Pacific GOS CNGrid China GARUDA GARUDA India OurGrid The OurGrid Community Brazil Recently, glite development was taken by the EMI 28 project, together with the consolidation of other European middleware, ARC and UNICORE, merging them in a unique European distribution. The release of EMI will be integrated by EGI in the Unified Middleware Distribution (UMD) and will eventually replace the separated glite, ARC and UNICORE distributions. The usage of infrastructures with different middleware is required by VRCs that are spread across several continents. Interoperability is thus intended not only between each infrastructure and UMD but more generally among all the combinations. This ambitious goal should be reached using standards developed or supported by Open Grid Forum (OGF) and identified from Grid Interoperation Now group of OGF as technologies enabling interoperability. EGI-Inspire and EMI projects are very active in this context. Some of the selected standards were developed by or in collaboration with European middleware editors. In the following paragraphs the middleware reported in Table 1 will be shortly described UMD The Universal Middleware Distribution (UMD) is provided by EGI with the aim to provide a complete set of software services for EGI's user and operation communities to use. Products offered by external technology providers such as EMI, IGE and SAGA comprise each UMD release GOS Presently at version 4.0, it is maintained by several institutions under the Chinese Academy of Sciences. It is strongly based on GLOBUS but completely re-written in Java with a quite different layout. The development of new versions is currently frozen and new directions for the future evolution are being evaluated /50

20 GARUDA It is based on GLOBUS WS and the meta scheduler GridWay 5.7. The resources information system is based on GLOBUS MDS, not on BDII. Certificates of the euindia VO are directly mapped on CEs gridmap file without the usage of VOMS. The data management is limited to gridftp in the head node of the CE, without gridftp clients in the Worker Nodes OurGrid It is a peer-to-peer computational grid targeted to run Bag-of-Tasks (BoT) applications on the idle cycles of organisations' desktops. It is based on three main elements: OurGrid Broker: it is the scheduler, usually installed on the machine of the user that is submitting jobs to the grid; OurGrid Peer: it is in charge of managing the desktops on a site, i.e. an administrative domain; it also allows the discovery and allocation of desktops that are available in remote sites; OurGrid Worker: runs in the desktops that are made available to the grid; it is in charge of identifying when the desktop can be used by the grid, and protect the desktop from any harm that could be caused by the grid jobs it executes Interoperability plan A realistic plan cannot be detailed without the agreement of the middleware developers involved. The key element is the usage of selected standards to grant some level of sustainability of the interoperability among the different infrastructures. The requirements to be satisfied are: To be able to send a job to another infrastructure; To be able, to some extent, to check the status of the job; To be able to retrieve the job output; To be able to manage and locate data. The OGF adopted standards that can be selected to implement a first version of interoperability are the following: Shibboleth and SAML to translate authentication and authorization assertion between different middleware; GLUE as a standard metadata schema for information system and BDII as unique container for the information from all the infrastructures; SAGA and OGSA-BES to translate requests of operations on jobs between different middleware; SRM for the data management; LFC as a catalogue for the files in the SRM shared between the infrastructures. In the following table, some reference documents are listed. 20/50

21 Topic AAI AAI BDII SRM LFC OGSA-BES SAGA Table 2 Reference documents Web Link From a VRC point of view, it is possible that the medium/long term solution of a true interoperation looks too far on the horizon and short term needs could be satisfied by a portal/gateway to directly submit jobs and handle data to different infrastructures but with a simplified environment. This is the use case of scientific gateways that have been recently developed by INFN Catania and used in other projects. Their implementation is anyway based on the use of standard solutions like SAGA. The next step is thus to agree with the middleware developers and maintainers a plan of implementation of the required missing components. Different technical solutions would affect the amount of human resources to be allocated and timings. The CHAIN recommendation is to continue to organise technical meetings with the relevant people involved to discuss deeper in detail the possible implementation and finalize the roadmap to the interoperability Specific Regional solutions Africa (sub-sahara) Regional coverage: This region covers all countries south of the Sahara desert. As mentioned before, the Sub- Saharan region is represented by two distinct realities. The region is generally under developed in terms of infrastructure in general and particularly in terms of e-infrastructure, while South Africa has recently made great improvements in this area. The development in South Africa is therefore conscientiously undertaken with the idea of proposing a blueprint to the other countries in the region, with which South Africa indeed has good scientific cooperation relations. The questionnaire was filled by Kenya, Ethiopia, Ivory Coast, Somalia, Nigeria, Sudan, DRC, Cameroon, Burundi, Zimbabwe, Malawi and South Africa. 21/50

22 Regional Grid state of the art resources, operations and middleware: Most of Africa is a greenfield region with isolated but growing patches of infrastructure. South Africa has a significant e-infrastructure deployment, which is mostly centrally coordinated by the Department of Science and Technology. This investment covers a supercomputing centre (the Centre for High-Performance Computing), and a national high-performance, low-latency network backbone, SANReN 29. Complimenting this centralised investment in infrastructure is a ground-up collaboration by several of the country's universities and national laboratories, in the form of a JRU. Complimenting this centralised investment in infrastructure is a ground-up collaboration by several of the country's universities and national laboratories, in the form of a JRU, which first started joint activities in The operational model was greatly influenced by the experience of the series of EGEE projects, but even more so by the close collaboration with the INFN and the experience of INFN Grid. Over the period , a national grid infrastructure was deployed, using existing resources at the partner institutes. These consisted of about 2,000 CPU cores and 10 TB of permanent disk storage, distributed around the country, which were integrated with the glite middleware stack. It should be noted that the NREN was only commissioned fully during the latter half of 2010, but nevertheless the training and dissemination activities were undertaken. Even though the performance of the network was far below the levels needed for a production national grid, the investment in time dedicated to training and technical deployment was willingly made by the institutes, thanks to the credible prospect of the deployment of the national network. When eventually the NREN was commissioned, the grid sites were ready and several of the first tests on the NREN were indeed made by grid applications. The other parts of the sub region are at different stages of development with those involved in other grid initiatives/projects such as the EPIKH and HP/UNESCO Brain Gain at relatively advanced stages. NRENs at organisational level exist in most of the region and working to develop the infrastructure and the same will host the NGI. Grid sites are available in Kenya, Ethiopia, Cote d Ivoire, Nigeria and Tanzania and have growing communities. There is growing interest in the Democratic Republic of Congo and Eb@leGrid is being set up with the main drive being to facilitate drug discovery from plants in the Congo forest. HPC facilities are also available in some countries, but are not interconnected. Zimbabwe is attempting to set up a local grid node but has bandwidth and expertise constraints Uptake of grid computing in sub-saharan Africa has been slow due the low level of interconnectedness of the region and high cost of connectivity. With emergence of NRENs over the past 5 years, and landing of undersea optical fibre cables along the East African coast, the situation continues to improve and more prospects lie ahead as the AfricaConnect project will begin to roll out the regional network in the coming months. This will enable the isolated grid sites to be interconnected. User communities and support: A main driving force behind the development of the grid in South Africa was the need of large scientific communities participating to international research collaborations. Already in 2003, with the adhesion of South African laboratories to the ALICE experiment at CERN, work began on development of coherent service deployment, albeit in an environment which was entirely lacking in infrastructural services. While a very large group of individual researchers exist, to a large extent unknown to each other and to the e-infrastructure initiatives themselves in the Sub-Saharan region, due to their incoherent nature, they do not constitute the driving force behind investment. Therefore, although not strictly-speaking representative, the large 29 SANReN South African National Research Network : 22/50

23 scientific communities supported in the region are useful, providing visibility and justification for e-infrastructure investment by the various institutions and governments in the region. In South Africa, these communities are: 1 HEP the ALICE and ATLAS experiments in particular; 2 nuclear medicine the GEANT4 collaboration; 3 structural biology the We-NMR collaboration; 4 drug discovery the biomed VO and the WISDOM grid collaboration; 5 Earth Observation the GENESI-DR project, and GEO (Group on Earth Observation). Apart from these, similar coherent, multi-institutional activities are being undertaken in the region, even if they are not explicitly dependent on e-infrastructure (but could nonetheless benefit greatly thereby). The most visible of these is the continent-wide interest in Astronomy, especially in Sub-Saharan Africa. The presence of several telescopes in the region, in the optical, radio and x-ray spectra (SALT/SAAO, HartRAO in South Africa, HESS in Namibia) form the basis for very strong scientific collaboration and exchange, while the prospect of the region hosting the Square Kilometre Array has provided an important consolidation of infrastructure and research policy in the region. Apart from these, projects in the natural products (exploiting e-infrastructure for therapeutic and cosmetic drug development and discovery), earth observation natural and resource management are also areas where consistent scientific collaborations exist and could easily and effectively exploit e-infrastructures made available to them. However, the overall goal of the national grid initiative was to reduce the barriers experienced by the majority of the entire region's researchers when accessing and using high-performance e-infrastructure. The principles of inter-institutional coordination and collaboration, relying on the best practices learned during national and regional initiatives across the world allowed for a national grid to be deployed in a comparatively short time in South Africa. This experience has been documented and promoted at several regional concertation and research conferences and continues, integrating experiences from regional projects such as EUMedGrid Support, EELA and SEEGrid. The key to success of the entire enterprise is acceptance and uptake in the user community, as well as to a lesser extent in the technical communities which need to support and extend the grid services and applications. The EPIKH project has provided the framework for training and knowledge transfer, while the HP-funded, UNESCO managed Brain Gain Initiative 30 has provided a critical mass of committed researchers, in a well-organised human network. What is more, the existence of formal scientific collaborations, with their associated software for data processing and analysis has given impetus to South African researchers to commit to learning how to use the SAGrid infrastructure. These have been particularly in the domains already well-known throughout the world (HEP, life sciences, chemistry, nuclear medicine), but nevertheless exclude the vast majority of potential users which exist in the Sub-Saharan region, let alone only South Africa. Despite the existence of a modest, but stable integrated infrastructure in South Africa, the so-called digital divide in Africa is felt so strongly that nonetheless most scientists are prevented or discouraged from using it. Due to the lack of experience and familiarity with the concepts of a distributed, service-oriented computing infrastructure, most users remain locked into a desktop-centric computing model; as such, institutes outside of South Africa are hesitant to commit to policy changes to open up access to what is perceived to be an extremely rare resource, represented by the network bandwidth and HPC hardware. Indeed, the truly rare resource is skilled manpower and support in the region. 30 BGI website : 23/50

24 Relationship with underlying networking activities: As mentioned earlier, though NRENs in the region have boomed in the past 5 years, apart from South Africa and Kenya, the networking infrastructure is yet to be developed. In some countries, the NGI is being developed and coordinated outside the NREN while in others the NGI fall within the NREN. In South Africa, where SAGrid is advanced, there has been a concerted effort to integrate the activities of the network with those of the national grid in South Africa. The reasons are organisational, as well as functional. Both SANReN and SAGrid are coordinated by the Meraka Institute, but directly involve staff and users at the various institutions; moreover, SANReN perceives the combined services provided by the grid as a set of advanced services which are enabled by the network, providing value to the user communities. Therefore, every effort has been made to streamline training and development activities within the country and in the region, between the network and distributed computing activities. In DRC the Eb@leGrid is being developed within Eb@le the NREN. During data collection, in the rest of the region, where NGIs do not yet exist, the NREN was asked to complete the questionnaire. Sustainability and NGIs: Due to the relative vacuum of activity, resources and support available in the region, a great effort was required to kick-start the development of the human resources and technical collaboration necessary to support the desired uptake of e-science. This was of course supported by the institutes which would eventually be providing the resources (both computing resources and human resources) which would be used by the various user communities. The collaboration of these institutes was formalised through the creation in South Africa of a Joint Research Unit, which is represented by the South African Council for Scientific and Industrial Research (CSIR). The coordination activities are undertaken by the CSIR, through its ICT-specific research institute, the Meraka Institute 31. The activities initially consisted essentially of technical training and dissemination activities, for the most part undertaken in collaboration with the GILDA training infrastructure and the INFN, in the context of the EGEE-III project. Great care was taken to ensure that knowledge was transferred to South Africa, in order for the activities in the country to become self-supporting, as well as to act as eventual base for diffusion of knowledge and support through the Sub- Saharan region. While this transfer has been achieved, the final stage for the development of the grid activities (and scientific exploitation thereof) is still under development. The plan is to see the creation of a legal entity which is responsible for the various activities already under way, and to act as a contact for funding agencies, thereby able to participate collectively to projects. Since such an activity is already under way in South Africa for the NREN, and due to the very close alignment of the network and distributed computing activities, the proposal is to merge these two into a true integrated infrastructure, providing valuable services over the network to the users, independent of their location. The sustainability and long-term funding of this entity is proposed to take the form of a centralised component coming from the national government and a co-funding contribution from the participating sites. The central component will fund activities which are undertaken in common, independent of the individual institutions, such as application and user support, training and so on, while the co-funding contribution will contribute to site resources, including human resources. While the development of the NGI document is under way, great care is taken to consult each participating institute, as well as those which could potentially participate, in order to have a representative solution /50

25 This procedure is of course proposed to the other countries in the region, however great care should be taken when considering the national realities of the region, where the single-ngi concept may not be relevant. The nature of the user community should be considered, since it is sparse and human mobility is low. Here, the experience of the Balkan states and to certain extent the North-African states is being used as a relevant benchmark Asia-Pacific Regional coverage: The region covers both North East Asia and South East Asia areas including Australia, Indonesia, Japan, Malaysia, Pakistan, Philippines, Singapore, South Korea, Taiwan, Thailand and Vietnam. Regional Grid state of the art resources, operations and middleware: International network connectivity is one of the key components of e-science infrastructure. The provision of network connectivity in the Asia Pacific region widely differs from country to country and among different institutes in the same country. At the moment, the Asia-Pacific Advanced Network (APAN) comprises network circuits provided by major member countries such as Japan, Korea and Taiwan, etc. and provides the widest geographical coverage in this area. In addition to APAN, the Trans-Eurasia Information Network (TEIN3) provides further connectivity, particularly to the less well-connected countries in South East Asia. In addition to these regional structures, some countries have their own connections to Europe and the US. In total, the bandwidth capacity from North East Asia (including China, Japan, Korea and Taiwan) to Europe and the US exceeds 50 Gb/s. In contrast, the connectivity within the South East Asia region is relatively weak. There is a wide range of Grid/Cloud technology penetration and e-science application in Asia Pacific region. The Grid related EU initiatives under the Seventh Framework Programme (FP7) continue to be the driving force behind the e-science collaboration in the Asia-Pacific region. AP e-science collaboration started from the Asia Federation in the Enabling Grid for e-science (EGEE) project. In 2008, the launch of EUAsiaGrid project has been the primary driver for the e-science infrastructure and Grid related developments in the region. Members from the Asia Federation and EUAsiaGrid, now forming the Asia Pacific Grid Initiatives (APGI), joined the Integrated Sustainable Pan-European Infrastructure for Researchers in Europe (EGI-InSPIRE) project and continued facilitating regional collaboration and bridged Asia with the world. The operation of these resources as a federated regional infrastructure linked in with other regional infrastructure is ensured through the site certification and monitoring supported by APROC, the regional operations centre, which also provides core services such as a virtual organisation management, a region-wide grid information system, monitoring systems and a workload management system. At the technical and operational level, we can see that through the work of EUAsiaGrid, an increasingly mature infrastructure is emerging that enables researchers in the region to engage in e-research collaborations with colleagues in other countries and in the wider world. The WLCG Asia Pacific regional operation centre, run by Academia Sinica Grid Computing Centre (ASGC), has been supporting 38 sites in 15 countries (Australia, India, China, Japan, 25/50

26 Korea, Malaysia, New Zealand, Pakistan, Philippines, Singapore, Taiwan, Thailand, Indonesia, Hongkong, and Vietnam. According to the Asia Pacific Grid Policy Management Authority (APGridPMA), there are more than 1,800 valid user certificates for the moment. Recently, ASGC has supported most sites to reach more than 90% of resource availability. From January 2011, there have been about 62,000 jobs per day, which is 10 times more than 3 years ago. CPU-hours have also increased by 10 times to 80,000 per day compared to January LHC experiments are the primary users in APGI, but the EUAsiaGrid project has also drove up the growth since it began in 2008, along with other distributed e-science collaborations in the region. User communities and support: EUAsiaGrid has been the major e-science project in Asia-Pacific region in the past three years. The table below describes various applications and projects that have been created and sustained through this partnership. Table 3 - e-science Collaborations in Asia Pacific Discipline Applications Partners High Energy ATLAS, CMS, ALICE, BELLE, CDF, TH, TW, CESNET, INFN Physics GEANT4 Biomedical virtual screening for drug discovery- MY, TW, VN, CESNET, INFN, Avian Flu, Dengue Fever. Pandemic FR disease analysis Bioinformatics Grid enabling phylogenetic inference, SG, TW, VN, CESNET, INFN SVN parameter optimization for prediction of caspases, Genome search to ID T3SS effect, autodock ligandreceptor docking, complex disease studies Earth Sciences Disaster mitigation for earthquakes ID, MY, PH, TH, VN, TW, CESNET, INFN Computational Chemical compound property analysis TH, TW, CESNET Chemistry Climate Change Weather simulations, sea level rising ID, TH, PH, VN, TW Social Sciences Social simulations TW, UK The major application areas are described as follows: High Energy Physics: The participation in the Worldwide LHC Computing Grid (WLCG) continues to be the incentive for many countries to advance the Grid technology. The HEP user groups in Asia-Pacific regions can be clustered according to their involvement indifferent projects as follows ATLAS: U. Melbourne (AU), U. Tokyo (JP), ASIoP (TW); CMS: Mindanao U. (PH), KNU(KR), TIFR(IN), NCP(PK), U. Canterbury (NZ), NTU (TW), NCU (TW); Belle: U. Melbourne (AU), KEK(JP), NTU (TW), NCU (TW); CDF: ASIoP (TW); Accelerators for Therapy : U. Melbourne (AU), Philippine Nuclear Research Institute (PH), Vietnam Atomic Energy Commission (VN), Chulalongkorn U. (TH), U. Malaya, 26/50

27 MINT (MY), U. Indonesia (ID), Nat l Cancer Center (SG), ASIoP (TW), NCU (TW), CGU (TW), NTHU (TW). Bioinformatics: Most of AP users focus on the use of grids to deal with some aspects of the new drugs discovery for diseases that are pandemic in the region. The GAP Virtual Screening Service, developed by ASGC, is becoming widely used as the environment for parametric and docking studies in this region. Vietnamese hospitals started to use HOPE Hospital Open Software platform for E-Health as their interface to the grid and also as the environment for collaboration with the French partners. Computational chemistry: The grid infrastructure has been successfully used as a high performance platform to study material properties. Also, the EUAsiaGrid project helped users from the Asia-Pacific region to get access to the Gaussian VO and through it to get access to the commercial licensed Gaussian package. Disaster mitigation: Similar to the Bioinformatics and e-health, these both areas are of utmost importance to the Asia-Pacific countries. The disaster mitigation focused primarily on the seismic wave propagation simulation and use of grid as a platform to widen and extend international collaboration. With the understanding of serious difficulty to predict an Earthquake, the work focused both on improvement of sensor network and on-demand generation of seismogram at any location, and also on the precomputation of probable scenarios to have them readily available in case of a real disaster. Social science: Grids represent a very promising tool to overcome restrictions and constraints imposed on access and processing of the sociological (census) data. Grids can not only provide sufficient storage and processing power for complex analyses, but their distributed nature supports processing models without central control. Queries can run over several census depots, with just anonymized results being available for researchers (as required on the limits on access to the primary data). Relationship with underlying networking activities: The regional network APAN (the Asia Pacific Advanced Network) refers to both the organisation representing its members, and to the backbone network that connects the research and education networks of its member countries/economies to each other and to other research networks around the world. APAN Ltd is the not-for-profit Association that is the legal entity created to undertake the activities on behalf of APAN members. APAN members are the entities representing research and education network interests in the countries/economies of Asia and Oceania. As mentioned earlier, AP region has been working closely with European e-science projects such as EGEE, EUAsiaGrid and EGI-InSPIRE. The regional network used to connect to Europe was mainly TEIN2. APAN and TEIN* Co-operation Center (TEIN*CC) signed a Memorandum of Understanding on 23rd August 2011 to collaborate on projects to benefit the research and education community in the Asia Pacific region. Sustainability and NGIs: In general, the Grid activities in the region are growing at different speeds. Many countries in the Asia-Pacific region have already laid foundations for their national e-infrastructures. For example, in Singapore National Grid Service Providers were commissioned by the Infocomm Development Authority to provide compute and storage resources based on the utility model. The Korean Institute for Science and Technology (KISTI) has been instrumental in development of operational e-infrastructures in Korea. The KnowledgeGrid initiative has a similar broad aim in Malaysia. 27/50

28 Central Asia Regional coverage: This region covers Kazakhstan, Kyrgyzstan, Tajikistan, Turkmenistan, Uzbekistan, and also Iran and Pakistan. Southern Caucasus - Armenia, Azerbaijan and Georgia are considered to be affiliated with the European region, where notably Armenia and Georgia are members of EGI. The questionnaire was filled in by Iran only, while Kazakhstan and Kyrgyzstan provided minimal answers. Regional Grid state of the art resources, operations and middleware: The region is entirely green-field, with some previous developments in Pakistan and Iran only. There are no major resources deployed, apart from Iran with 7 Grid sites and 650 CPUs. A supercomputer is available in Kazakhstan. No operational solutions are deployed, apart from CA in Iran User communities and support: User communities are very limited, with main candidates being Seismology, Weather Forecasting, Mechanical Engineering. Relationship with underlying networking activities: The CAREN network is established in 4 out of 5 central Asian republics, thus providing a good base for future Grid collaboration with Europe. Sustainability and NGIs: No NGIs are available apart from Iran and no regional solution is available even on the conceptual level China Regional coverage: This region covers Chinese mainland, Hongkong SAR, Macau SAR. Regional Grid state of the art resources, operations and middleware: There are two Grid infrastructures in China i.e. China National Grid and China Education Grid. China National Grid (CNGrid) is the largest and the most active one in this region. It has aggregated computing power of the super clusters including former No.1 supercomputer in the world from more than 10 supercomputing centers including one located in Hongkong SAR. China National Grid uses their own middleware called CNGrid GOS and monitoring system called CNGridEye. The whole environment is sustained by China National Grid Operation and Management Center. There is no Grid node in Macau SAR yet now. User communities and support: User communities are quite large in China. There are more than 1,500 registered group users now to use high-performance computing capabilities via China National Grid. Also, China National Grid has established two domain user societies (i.e. industry simulation domain and biological information domain) and one general user society to support different kind of users. 28/50

29 Relationship with underlying networking activities: The Grid infrastructure between China and Europe has already been connected via several different networks (TEIN and ORIENT). Hopefully, it may be faster via ongoing networking activities. Sustainability and NGIs: In next five years, China will upgrade the scale of China National Grid environment on either the computing power or the covered area or both. The infrastructure will support more domain applications via traditional high-performance computing service or some new service styles such as cloud and grid society India Regional coverage: The region of interest is India that is also a country of Asia, but due to its large territory and economical impact, it s always considered a sub-continent. Regional Grid state of the art resources, operations and middleware: As indicated in section 3.1 two main grid infrastructures operate in India: The GARUDA India National Grid Initiative and the regional component of Worldwide LHC Computing Grid (WLCG). Both initiatives closely cooperate with the EU-IndiaGrid and the CHAIN projects. Figure 9 Map of GARUDA GARUDA initiative is a collaboration of science researchers and experimenters on a nationwide grid of computational nodes, mass storage and scientific instruments that aims to provide the technological advances required to enable data and compute intensive science of the 21st century. One of GARUDA s most important challenges is to strike the right balance between research and the daunting task of deploying that innovation into some of the most complex scientific and engineering endeavours being undertaken today. 29/50