Medical Bioinformatics and e-Bioscience

e-BioInfra Advanced Page

Here you find more detailed information about the e-BioInfra software and services (see fig).

Grid and Middleware Services

Starting from the lower level (red layers in figure above), the computing and storage resources available at the Dutch Grid are exposed as "computing elements" (CE) and "storage elements" (SE). Other gLite middleware services for management of jobs, storage and files, facilitate the access to these computing and storage resources via an abstraction layer. For example, instead of submitting a job to a particular computer cluster of the infrastructure, the system simply submits a job to "the grid" and the gLite middleware takes care of scheduling the job in the most optimal way (e.g., in the least loaded cluster). Likewise, gLite's file management services can handle replicas and load balancing for data transfers (e.g. distributing access among the existing replicas). Additionally, a San Diego Storage Resource Broker (SRB) is available for storage of data and metadata.

By using the gLite middleware to build the system, the e-BioInfra is based on a scalable platform that is compatible with continued developments in the EGEE project.

VBrowser

Now going all the way to the upper level (green layer in figure above), the user communicates with the system with the Virtual Resource Browser (VBrowser). This is a program written in Java that implements access to local and remote resources (data and services) from a user friendly graphical user interface (GUI). It runs on Windows, Linux and Macintosh platforms, requiring minimal software installation and configuration at the user's workstation. (The only requirements are a Java 1.5 installation and a network connection.) From the VBrowser, the user can directly manipulate files and run workflows on the grid. For example, to interactively copy and paste files from the local computer to some grid storage.

More details about how to download, configure and use the VBrowser are presented in the VBrowser page.

e-BioInfra Services

Currently three services are available to intermediate the communication between the VBrowser and the grid middlware (blue layer in figure above). These services handle the technical details and implement (very simple) fault tolerancy mechanisms to facilitate usage of the Dutch grid. Accordingly, three plugins were developed for the VBrowser to implement interfaces with these services:

* The MOTEUR plugin/service: MOTEUR is a workflow management system used to execute workflows on the grid. For each component of the workflow, MOTEUR submits a job on the grid using the glite middleware. More details about MOTEUR are presented in the MOTEUR page.This service is installed in the webserver.

* Job Monitoring plugin/service: obtains the status and outputs of jobs executed on the grid - see more details in the Job Monitor page.

* LFC plugin/service: implements access to files on the gLite Logical File Catalog. See more details in the LFC page.

Note that these services encapsulate much of the necessary steps to access the grid, and avoid the necessity to install heavy software packages on the user's workstation (such as the gLite middleware).

Usage in a nutshell

Figure 3 - Overview of e-BioInfra usage from the VBrowser

Documentation

General Information:

• GettingStarted
• SARA's Grid Wiki
• EGEE's Users Wiki
• How to get a grid certificate
• How to become a member of a Virtual Organization * e-mail list for e-BioInfra users: archives, membership page

Software Tools

• Virtual Resource Browser (VBrowser)
• Workflow management with MOTEUR
• Generic Application Service Wrapper (GASW) and the stand alone GASW utility SaGasw
• Miscellaneous scripts and tips
• How to run matlab jobs on the grid
• Tips and handy scripts to manipulate grid files

Running experiments on the grid

• Grid accounting portal
• From executable to workflow