Integration Fest 4

Scope and Purpose

The purpose of IF-4 is to integrate myGrid components into a myGrid system that can be used to demonstrate myGrid at the following meetings:

• ISMB 2003, Brisbane Australia, 29 June - 3 July 2003
• UK w-Science All Hands Meeting 2003, Nottingham, UK, 2-4 September 2003

The demonstration will illustrate use of myGrid in the context of the GravesDisease scenario.

Time and Place

IF-4 will take place from Monday to Friday, 2-6 June 2003 in the Mixed Reality Laboratory at the Department of Computer Science, Nottingham University.

We have booked accommodation at [[http://www.innkeeperslodge.com/lodgedetail.asp?lodgeID=45][Innkeeper's Lodge, Derby Road, Wollaton Vale, Nottingham]].

Agenda

Currently, the only fixed events during the week are:

• Monday, mid-morning - lunchtime: arrive, set up working environment
• Friday (with investigators): demonstrate integrated system, script and record Graves' Disease demonstration, identify remaining problems

During the week, we will also take the opportunity to update some of the higher-level descriptions of myGrid and its components on our website.

Preparation

To help us use our time more effectively during IF-4, the Newcastle team have offered to build and test myGrid components at Newcastle in advance of the week in Nottingham. To do this, we need to make components available to them and the rest of the myGrid team in advance of 2 June.

At the investigators' meeting at Sheffield, 8 May 2003, we proposed the following delivery schedule:

By 'delivery', we mean:

• changes committed to CVS
• buildable
• runnable

Each component should contain a readme file (HTML or text) to describe how to deploy and run the component. This file should identify any configuration parameters/properties that may need setting or changing, and any external components that need to be present (e.g. Tomcat servers and database systems).

If the build process differs from the myGrid ModuleBuild conventions, those differences should also be described there.

Suggested Schedule

This schedule mainly covers integration activities, where we initially install services or integrate several services. The GANTT chart is a finger-in-the-air guess of times, dependencies and the people involved at each stage. Since the Workbench is very much the focus for a demonstration, Kevin in particular will be doing several things at once - I hope as shortish periods of consultancy.

• GANTT Chart for IF-4:
  

1. Set up environment

Install tools, download sources

2. Local bioservices

We should try to install at least EMBOSS services locally, in case of network failures at ISMB

3. Install DB2 & MIR

We have these on Nottingham's Solaris machine. We need to chaeck that we can get it and the MIR onto PCs for the demo.

4. Load test data & workflows

All the Graves' disease inputs and workflows; other available stuff.

5. Base Workbench & MIR

To give ability to browse the work context 'filesystem'.

6. Independent workflow enactment

Invoked from outside the Workbench

7. Integrated workflow enactment

Enactment of MIR-stored workflow from workbench (when the user knows where the workflow is).

8. Notification service

Install and test of basic service.

9. DB Notification service

Install and test of basic service.

10. Notifying MIR updates

Integrate DBNS with Notification service, MIR and Workbench. Workflow status is notified indirectly (by trigger on matching update to MIR).

11. Independent workflow description

Install and test Pedro-based service & workflow description& view service. These need not be integrated with the Workbench: description is currently an off-line activity.

12. Integrated workflow description

Integrate Pedro-based client into Workbench ONLY if it turns out to be easy.

13. Workflow discovery

Integrate Workbench wizard, Find service and View service

14. Workflow display

Graphical read only view in Workbench

15. Workflow edit

Edit tool (graphical?) in Workbench, possibly integrating with service discovery.

16. Independent DQP

Install and test of basic service.

17. Integrated DQP

Search over three MySQL^?-based services (GO, Affymetrix mapping DB, ArrayExpress^?). Will wrap DQP in a Web Service and invoke from a workflow.

18. Text Extraction

Invoke TE service from workflow. Depends on Neil and Guo making TE results accessible to Java clients.

19. MIR queries

Results of 'interesting' queries in Workbench

20. Provenance display

Based on existing XSL transformation.

21. Provenance queries

Combine provenance information and MIR contents to produce more interesting 'traces' of inpu/output relationships?

-- NickSharman - 23 May 2003