Community Forum 2012: Training workshops - go with the (work)flow
John Walsh introduces the community training workshops
The Community Forum 2012 will see experts across the scientific domains gather to both publicise new results from their grid-based computational models, and to display the wealth of tools available to help experienced grid users and unversed scientists to exploit the EGI distributed computing infrastructure (DCI) with greater ease.
The Heavy User Communities (HUC) training sessions and workshops shall be as diverse as ever. In addition to a range of Tools and Shared Services workshops, this year will also see a greater focus on Workflows and Portals.
Here is a preview of the up-and-coming workshops.
Workflow Management Workshops
Workflows describe a set of linked tasks or processes that yield an outcome. Scientific workflow management systems are particularly good at handling a series of complex tasks using large datasets in numerous data formats. A workflow can be described once, but used many times with different input parameters. Portals, on the other hand, can provide user-friendly, web-based access to workflow management and job submission systems. Indeed, Workflow Management Systems and Portals, play a huge role in managing scientific workloads across all the scientific disciplines, and are expected to have a greater role to play in our daily lives as more and more people move to tablet-based computers.
Kepler
The Kepler Scientific Workflow Tutorial will be ‘hands on’, with attendees being introduced to: basic workflow concepts (e.g. frameworks, actors, directors); how to build basic workflows; how to use relations, paths and synchronization, if-else, loops; and finally, how to use advanced features, such as grid actors, in simple and complex workflow scenarios.
WS-PGRADE
The two-part tutorial on WS-PGRADE/gUSE presents an open-source, web-based workflow management tool that enables users to exploit several grid systems, use multiple grid technologies, and to access other DCI infrastructures. It has a feature-rich and easy-to-use interface. The first session focuses on User and Application development training and will cover: WS-PGRADE workflow concepts; parameter study features and support in workflows; the WS-PGRADE portlets; access to other DCI infrastructures; scientific application areas using WS-PGRADE. The second session will be at a more advanced level and covers WS-PGRADE/gUSE portal development. It is primarily aimed at scientific communities, NGIs or portal developers who would like to create customised WS-PGRADE based science gateway applications for their users. This tutorial will introduce the attendees to the Application Specific Module (ASM), a flexible API layer to assist the creation of customised science gateways.
SHIWA
This tutorial introduces SHIWA (Sharing Interoperable Workflows for Large-Scale Scientific Simulations on Available DCIs), a tool aimed at handling a wide variety of scientific workflows in a platform neutral way.
SHIWA allows workflows described in one format to be translated into other formats, and enables workflows to run on several DCIs. SHIWA also provides a searchable workflow repository, thus allowing e-scientists to reuse them in their own workflow applications.
This is important because communities that have developed workflows may be susceptible to vendor ‘lock-in’ – i.e., prevented from using them in other workflow management systems, and tied into particular computing infrastructures.
Tools and Shared Services Tutorials
The Tools and Shared Services tutorials look at a range of tools that can be exploited by all grid users.
CERN VM FS (CVMFS)
The CERN Virtual Machine File System (CVMFS) is a caching, http-based, read-only, scalable file system optimised for delivering experiment software to (virtual) machines. Other primary features include: standards-based technologies; file system security (SHA1 check via HTTPS); file-based duplication; pull-based solution for deploying experiment software at each site.
This session is aimed at user communities wanting to explore the use of CVMFS as a new, efficient and flexible means to manage and deploy their experiment software on the grid. In particular, the session will discuss the server-side configuration, and should be of interest to resource centre administrators and user communities alike.
Parallel Computing Workshop
This workshop (which is included in the 'Making DCIs work for you' session) will concentrate on the current state of the art for successful parallel job submission on the infrastructure. The topics covered will include: description and submission of parallel jobs; techniques for successful job submission, and pitfalls to avoid; execution of MPI applications using gLite, ARC and Unicore middleware stacks; support for other parallel application frameworks (OpenMP and hybrid MPI/OpenMP); use of GPGPUs.
StratusLab
With over 50 attendees, the StratusLab tutorial was one of the success stories at the EGI User Forum 2011. Participants will learn about cloud technologies in general, and discuss the distinction between Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS) Infrastructures. The workshop will discuss how infrastructures based on the StratusLab distribution can be integrated with EGI and how cloud services complement grid services. Practical exercises will teach the participants how to launch virtual machines, customise their computing environment, share those environments with others, manage virtual disks, and define complete services.
More information
This issue
