Suhaimi Napis, PhD Chief Technology Officer (CTO) and Director, Chief Technology Officer (CTO) and Director, InfoComm Development Centre (iDEC) Universiti Putra Malaysia suhaimi@putra.upm.edu.my, s.napis@gmail.com EGI-iNSPIRE (European Grid Initiative) JRU Partner, EU FP-7 Project Deputy Chair, Research Steering Committee and Head, E-Research WG, Malaysian Research and Education Network (MYREN) Chair (Interim), e-Research WG, Asia Pacific Advanced Network (APAN) Vice Chair, ASEAN Science and Technology Research and Education Network Alliance (ASTRENA) EUAsiaGrid Project Partner (Completed)
Agenda � MYREN Topology and Recent Development � On Demand Infrastructure for Research and Education Collaborative Technologies � MYRENCLOUD and Future Plans MYRENCLOUD and Future Plans
3/24/2011 3/24/2011
TEIN3 TEIN3 – Launched January 2009 Launched January 2009
Infrastructure Update Infrastructure Update International Cooperation and International Cooperation and Linkages Linkages MYREN Phase 2 Launched on 29 April 2010 MYREN Phase 2 Launched on 29 April 2010 KOREN (KR) KOREN (KR) North America North America CERNET (CN) CERNET (CN) JGN2 (JP) JGN2 (JP) PregiNet (PH) PregiNet (PH) SingaREN (SG) SingaREN (SG) Inheren (IN) Inheren (IN) AARNet (AU) AARNet (AU) Linking 33 Linking 33 countries in Europe, countries in Europe, 10 10 countries in Asia countries in Asia-Pacific and Pacific and Internet Internet2, the United States , the United States Creating a gateway for global research collaboration for over 60 Creating a gateway for global research collaboration for over 60 million scientists, academics and students. million scientists, academics and students.
E-Science Activities � Computational Sciences � Bioinformatics and Computational Biology � Cheminformatics and Computational Chemistry � Agrobioinformatics (Precision Farming), � Natural Resources Informatics (Biodiversity) � High Energy Physics and Nuclear Simulation � Engineering eg CFD using Open-FOAM � Grid and Cloud Computing � E-Culture (Digital Culture and Heritage)
(On Demand Infrastructure for Research and Education Collaborative Infrastru cture) � Video Collaboration Services � Individual or Group virtual meetings, hassle-free video conferencing systems including ubiquitous desktop video collaboration with recordings desktop video collaboration with recordings � 158 concurrent sites (60 in Full HD), even Laptops � 1VideoBridge, Northern Light, AccessGrid, EVO, etc � uCAST Video Sharing Portal – UPM’s “YouTube” � Streaming media facility � On Demand Rich Media for Learning
Eg :Rich Media for E-Learning � Recorded video collaboration sessions as VOD materials (Unedited and edited) � Plan to acquire Video Centre for recording and streaming to complement MCU � Over 6 years experience and using many technologies technologies � Challenges:Constantly in need of assistance from technical staff (costly) � Using simple readily available devices: webcams, built-ins or detached, handsfree personal speaker microphone, Xacti and Flip � EMPOWER the lecturer to do it by themselves
� Data Services � Data Fabric system for storing research data � Data Transfer Service for transferring research data � Database hosting services for collaborative research efforts research efforts � Virtual Server procurements � Centralised provision of development servers; no need for researchers to separately buy small servers; � Security Services � Malaysia Identity Federation and Access management (MyIFAM)
� Web Collaboration Services � Content collaboration and development � MYREN Web Collaboration Services MYREN Web Collaboration Services (hosted Twiki, Sakai, Plone) � Cloud Computing Services ○ High Performance Computing On-Demand over Private Cloud
BIRUNI GRID: HPC in Private Cloud BIRUNI GRID: HPC in Private Cloud • UPM has UPM has commisioned commisioned BIRUNI GRID with the hope BIRUNI GRID with the hope of of contributing HPC contributing HPC Clusters for Clusters for A-Grid Grid • Hardware Hardware Racking and Power up were done by Racking and Power up were done by suppliers while the rest of the configuration and suppliers while the rest of the configuration and deployment was done by UPM Grid Team deployment was done by UPM Grid Team • • Excellent Excellent learning opportunities for the Excellent learning opportunities for the Excellent learning opportunities for the staff learning opportunities for the staff staff staff • Many Many lesson learned; some are good and some are lesson learned; some are good and some are not so good but in the end the staff learned by doing not so good but in the end the staff learned by doing and get their training the hard way. and get their training the hard way. • UPM UPM get help from Academia get help from Academia Sinica Sinica Grid Computing Grid Computing and other international partners and other international partners
BIRUNI BIRUNI GRID GRID • BIRUNI GRID consists of three blade BIRUNI GRID consists of three blade-based clusters: based clusters: – Khaldun Khaldun Sandbox Cluster ( Sandbox Cluster (10 10 worker nodes with GB) worker nodes with GB) – Razi Razi Cluster ( Cluster (28 28 worker nodes with GB) worker nodes with GB) – Haitham Haitham Cluster Cluster (8 8 worker nodes with worker nodes with Infiniband Infiniband) • • Khaldun Khaldun Sandbox Cluster is used for Experimental Grid Khaldun Sandbox Cluster is used for Experimental Grid Khaldun Sandbox Cluster is used for Experimental Grid Sandbox Cluster is used for Experimental Grid (Development, and Learning purposes) while (Development, and Learning purposes) while Razi Razi and and Haitham Haitham Clusters are designated as Production Grid. Clusters are designated as Production Grid. • At any time, the At any time, the Haitham Haitham Cluster and Cluster and Khaldun Khaldun Sandbox Sandbox Cluster worker nodes can support Cluster worker nodes can support Razi Razi Cluster worker Cluster worker nodes on nodes on demand demand • High Performance Computing facilities for Computational High Performance Computing facilities for Computational sciences and Open sciences and Open-Source Source 3D Rendering Farm for D Rendering Farm for animation and screen rendering animation and screen rendering
BIRUNI GRID BIRUNI GRID-cont cont
Present and Future Present and Future Plans Plans • Grid Grid Application deployment for Bioinformatics/ Application deployment for Bioinformatics/ Cheminformatics and Medical Cheminformatics and Medical Informatics (Radiology) Informatics (Radiology) • Eg Eg: BLAST , MUSCLE, : BLAST , MUSCLE, ProDist ProDist (sequence analysis), sequence analysis), AUTODOCK (molecular docking), GROMACS (molecular AUTODOCK (molecular docking), GROMACS (molecular dynamics), dynamics), Mr Mr Bayes Bayes, BEAST (Bayesian phylogeny), GATE , BEAST (Bayesian phylogeny), GATE (Grid Analysis Time (Grid Analysis Time-Series Expression) , etc Series Expression) , etc • • Moving Moving from compute intensive to RAM intensive Moving from compute intensive to RAM intensive Moving from compute intensive to RAM intensive analysis from compute intensive to RAM intensive analysis analysis analysis • Next Gen Sequencing; Genome comparison Next Gen Sequencing; Genome comparison • Eg Eg: 7.5 TB raw data per run and 200GB of sequence : 7.5 TB raw data per run and 200GB of sequence data! >>Need MYREN to move data data! >>Need MYREN to move data • Other apps in the pipeline… Other apps in the pipeline… • Graphic/Image Screen Rendering ( Graphic/Image Screen Rendering (Opensource Opensource: Blender) : Blender) • CFD using Open CFD using Open-FOAM: need help from researchers of FOAM: need help from researchers of engineering departments engineering departments
MYRENCLOUD � MYREN 2 is set to offer MYRENCLOUD services � Video Collaboration � DataCentre/Disaster Recovery Centre � Rapid Server Procurement (1000 virtual servers (2XQuad, 8Gb RAM, 400Gb HD) servers (2XQuad, 8Gb RAM, 400Gb HD) � Virtual Storage (cloud drives): 100Gb for 10,000 users (researchers/educators) � Extending MYREN 2 to researcher centres of all ministries, users at schools, hospitals, libraries, community centres and others
Recommend
More recommend
