A Catch-All Science Gateway Portal for Malaysia Research

A Catch-All Science Gateway Portal for Malaysia Research Communities The International Workshop on Science Gateways 2013, 3 -5 June 2013 Elizabeth Pek Iee 1, Mohammad Yaser Shafazand 2 and Muhammad Farhan Sjaugi 3 1 NVG Scientific Sdn Bhd, 2 Universiti Putra Malaysia, 3 Academic Grid Malaysia 1 echia@novaglobal. com. sg, 279. zand@gmail. com, 3 farhansj@academicgrid. my

Outline • • Introduction Problems Professional Linux & Windows HPC Solution Services HPC & Cloud Infrastructures & Services Methodology Applications Portal & Automated Screen Shot &Fact Sheet Workflows Summary Computational Science – Compute Generated Imagery (CGI), Materials Acknowledgement Science & Bioscience, Numerical Weather Demo Prediction (NWP) Nova. Global Pte Ltd

Introduction • In Malaysia, the National Grid Initiative/Infrastructure is called Academic Grid Malaysia. Professional Linux & Windows HPC • The Academic Grid Malaysia is an open access Services distributed computing infrastructure Open to any HPC & Cloud Infrastructures & Services Malaysian Academia from institutions and organizations that are recognized by respective ministries to be Applications Portal & Automated involved/ associated in/with Distributed-, Grid-, Cloud- Workflows and Emerging Computing Applications/Projects Computational Science – Compute • The Academic Grid Malaysia provides opportunities to Generated Imagery (CGI), Materials the user communities to run their existing applications Science & Bioscience, Numerical Weather Prediction (NWP) more quickly and efficiently, and also to create ambitious new applications, without investing in extra hardware, and software resources. Nova. Global Pte Ltd

Problems • However, converting existing applications into Grid applications, and also creating new applications, currently Professional Linux & Windows HPC requires significant Grid knowledge, which many users do Services not have. HPC & Cloud Infrastructures & Services • Most of the user communities that have used Grid are accessing the resources via Linux command line. Applications Portal & Automated However some of the users are not familiar with Linux Workflows command line, moreover it becomes “extra” work for Computational Science – Compute them after their actual research work. Generated Imagery (CGI), Materials • This is one of the reasons of low Distributed Computing Science & Bioscience, Numerical Weather Prediction (NWP) Infrastructure (DCI) adoption among research communities in Malaysia. Nova. Global Pte Ltd

Solution • • • In order to solve the problems, we developed a catch-all science gateway portal as the “medium” to access the Professional Linux & Windows HPC Academic Grid Malaysia infrastructure. Services Catch-all here means the portal is not specifically developed/designed for a specific science application. The HPC & Cloud Infrastructures & Services portal is generic and is able to host many science applications. Applications Portal & Automated By providing science gateway portal, it is hoped that user Workflows communities no longer face problems and there is a common set standard in accessing the DCI (i. e. NGI Malaysia). Computational Science – Compute Generated Imagery (CGI), Materials Science & Bioscience, Numerical Weather Prediction (NWP) Nova. Global Pte Ltd

User community and scientific area targeted by the Science Gateway • There are three major user communities that are targeted by the science gateway: Professional Linux & Windows HPC – Life science is still one of the main research focus in Malaysia, hence many researchers are doing research on that particular field (i. e. Computational Services – – Biology and Chemistry). However, most of them are still using their own personal workstation to do simulation that could take days to complete. HPC & Cloud Infrastructures & Services Some of the reason why they have yet to adopt the DCI approach is because it requires resources which at this point of time they do not have. Applications Portal & Automated Creative industry (i. e. Multimedia) is one of the area that is actively supported Workflows by the government of Malaysia. Hence many academic institution offers programs in this field of study. However, some activities require decent Computational Science – Compute hardware for rendering and not all academic institution can afford this Generated Imagery (CGI), Materials requirement. Science & Bioscience, Numerical Weather Engineering domain generally has adequate resources in running their Prediction (NWP) simulation/application in High Performance Computing. However, most of the users rely on commercial/licensed Engineering software which is costly and has some limitation in the number of users and hardware (i. e. servers). Nova. Global Pte Ltd

Methodology • Long before the science gateway for Academic Grid Malaysia is available, NVG Scientific have developed a Professional Linux & Windows HPC portal consisting of Bioinformatics application portlets that Services is currently hosted at the National University of Singapore. HPC & Cloud Infrastructures & Services • However the portal is not suitable to be implemented at the Academic Grid Malaysia infrastructure because most Applications Portal & Automated of the portlets was hardcoded to work with the LSF Workflows software. Computational Science – Compute • Instead, we took the design of each portlet and re-write it Generated Imagery (CGI), Materials according to WS-PGRADE/g. USE+ASM framework. Science & Bioscience, Numerical Weather Prediction (NWP) Nova. Global Pte Ltd

Methodology • The segregation of the presentation and workflow layer in WS-PGRADE/g. USE framework gives us some benefit: Professional Linux & Windows HPC Services 1. The web programmer can focus on developing/coding the portlet HPC & Cloud Infrastructures & Services 2. The workflow designer can focus on developing/creating the application workflow that is the actual “workhorse” of the portlet. Applications Portal & Automated Workflows • The workflow shall be tested and validated by the Computational Science – Compute workflow designers and the end-users (volunteer) from Generated Imagery (CGI), Materials each user communities before the web programmer Science & Bioscience, Numerical Weather “glues” the presentation layer and the workflow together. Prediction (NWP) • After the web programmer glued the presentation layer and the workflow together, the portlet goes to the second validation test before it is released to the public Nova. Global Pte Ltd

How does it works? Academic Grid SCIBUS PORTAL Workload Management System Submit job Information System query Retrieve Professional Linux & Windows HPC status & output Services create credential HPC & Cloud Infrastructures & Services query Applications Portal & Automated Submit job Workflows File and Replica Catalog publish state Retrieve output Computational Science – Compute Job Logging status Generated Imagery (CGI), Materials Site X Science & Bioscience, Numerical Weather Prediction (NWP) Computing Element Storage Element Job status process My. Proxy + VOMS Logging and bookkeeping Nova. Global Pte Ltd 9

Screenshot Nova. Global Pte Ltd

Fact Sheet • • • Portal homepage : http: //scibus. cict. utm. my Portal Framework : WS-PGRADE/g. USE version 3. 5. 5 Professional Linux & Windows HPC Programming API : ASM 3. 4 Services Operating System : Scientific Linux 6. 2 HPC & Cloud Infrastructures & Services Grid Middleware : EMI Middleware UI version 2. 0 Applications Portal & Automated Applications: Autodock 4, AMBER, NWChem, GROMACS, Workflows Fast. DNAml, Phase, Clustalw-MPI, BLAST* , Muscle , Open. FOAM , Blender*, Multidendrograms, PROSPECT, NAMD*, Glimmer* Computational Science – Compute Generated Imagery (CGI), Materials Science & Bioscience, Numerical Weather * Will be released soon. Prediction (NWP) Nova. Global Pte Ltd

Summary • The NGI Malaysia provide opportunities to the user communities to run their existing applications more quickly and Professional Linux & Windows HPC efficiently, and also to create ambitious new applications, Services without investing in extra hardware, and software resources. • Most of the user communities that have used Grid are HPC & Cloud Infrastructures & Services accessing the resources via Linux command line. Applications Portal & Automated However this becomes a problem since not all of the users has Workflows Linux skills. This is one of the reasons for low Distributed Computing Infrastructure (DCI) adoption among research Computational Science – Compute communities in Malaysia. Generated Imagery (CGI), Materials By providing them a science gateway, it is hoped that user Science & Bioscience, Numerical Weather communities no longer face any problem in accessing the DCI Prediction (NWP) (i. e. Academic Grid Malaysia) and also increase the adoption of the DCI technology among research communities in Malaysia • • Nova. Global Pte Ltd

Acknowledgement • Herewith we would like to thank you: Professional Linux & Windows HPC – SCI-BUS project for providing us research grant to develop the Services Science Gateway for Research Communities in Malaysia. SCIBUS is supported by the EU FP 7 Capacities Programme under HPC & Cloud Infrastructures & Services contract n°RI-283481. NVG Scientific Sdn Bhd has been appointed Applications Portal & Automated as one of SCI-BUS subcontractor. Workflows – Centre for Information and Communication Technology – Universiti Computational Science – Compute Teknologi Malaysia for providing us resource to access to the Generated Imagery (CGI), Materials Malaysia Research and Education Network. Science & Bioscience, Numerical Weather Prediction (NWP) Nova. Global Pte Ltd

Demo • • • Login to : http: //scibus. cict. utm. my Username: demo@novaglobal. com. sg Professional Linux & Windows HPC Password: 45 scibus 123 Services HPC & Cloud Infrastructures & Services Applications Portal & Automated Workflows Computational Science – Compute Generated Imagery (CGI), Materials Science & Bioscience, Numerical Weather Prediction (NWP) Nova. Global Pte Ltd