Talisman rapid application development for the grid

1
Talisman - rapid application development for the
grid

• Thomas M. Oinn

Presented by Leighton Ing
2
Introduction

• Recall Process Interoperation lecture
• Ubiquitous computing(one user, many computers)
• Grid computing
• What is Grid Computing?
• As described by IBM
• Virtualization of distributed computing and data
  resources which presents a seamless, unified
  virtual computer to interact with

3
Why Talisman?

• Current package environment
• Powerful and expressive tools with a high level
  of complexity
• Goal Easy to use system which retains ability to
  perform complex analysis
• Enter Talisman...

4
What is Talisman?

• Application framework for building custom
  applications for specific needs which maintain
  ease of use and provide narrow functionality
• Production of a canned web-based application
• Runtime system for these applications

5
Talisman

• Basic installation

6
Talisman Definition Language

• XML
• Define the application behavior in terms of data
  encapsulation and actions
• Does not contain formatting information
• Assembled into a server side data structure
• Presented to the user as a set of Web pages which
  the user can interact with in the ways described
  by the definition

7
Talisman Definition Language

• Includes a set of core data types and actions
• Fields and triggers
• Extensibility mechanisms
• Bridging code for interoperation with novel
  systems
• Example use of Martin Sengers SOAP web service
  wrapper around the EMBOSS tool set Written in 2
  days
• Tree structured definition system
• Arbitrary levels of subpages

8
Provenance Tracking

• Function provided by the runtime environment
• Provides a documentation trail of the session
• Results and metadata
• Vital in grid computing
• Multipurpose
• Easier debugging of new applications
• Management reports
• Lab reports

9
Provenance Tracking

• Easier debugging
• Inspection of a session for proper operation of
  the application
• Management reports
• XML document output can be parsed to generate
  reports on user activity
• Lab reports
• Automated recording of what has been done, for
  storage or dissemination

10
Provenance Tracking

• Type of information recorded
• What did I do?
• When did I do it?
• How was the process run?
• Metadata about the applications accessed and the
  parameters involved e.g database versions
• Who ran the process?
• Why was a particular choice made?
• On the fly annotation of the interactions

11
Plugins

• RDBMS access
• Read/write access to data in relational database
  management systems
• Login data, connection pooling, auditing, other
  basic functionality
• Web, .net, Grid servides
• Invocation of web based services called from
  Talisman
• Some assembly required
• Examples BLAST, Bibliographic Query Service

12
Plugins

• Ensembl integration
• Definition files which allow access to searches
  and data retrieval
• Bioinformatics service access
• EMBOSS functionality via the SOAPLAB package
  written by Martin Senger
• Assist workflow building
• MyGrid component interaction
• Web interface to higher level MyGrid functions

13
Talisman Runtime Environment

• Presents a web page front end to the user
• No need for specialized client installation
• Minimal client requirements
• Hot deploy and update of applications
• Application updates simple no restarts
• New versions available to users the next time
  they start it
• Minimized downtime

14
Talisman Runtime Environment

• Persistent server state
• Suspension and recollection of sessions
• Permits user to save and load their project as
  desired
• Fault tolerant clustering through a third party
  distributed session manager possible
• Platform Independence
• Java servlet support required
• Tested on Win32, i86 Linux, MacOS X, Solaris

15
Talisman Runtime Environment

• Open standards
• Talisman licensed under GNU General Public
  License(GPL)
• Built with software also under GPL
• Usable on commodity hardware, open source
  operating systems and server software

16
Interaction through Taverna

• Use Talisman as a distributed service through the
  Taverna workflow system
• Tune in next time for the presentation on Taverna

17
Availability

• Test installation at EBI
• http//www.ebi.ac.uk/collab/mygrid/service1/talism
  an
• Some functionality not available
• Project page at SourceForge
• https//sourceforge.net/projects/talisman
• Anonymous CVS access

18
References

• http//talisman.sf.net
• http//www.ebi.ac.uk/collab/mygrid/service1/talism
  an/index.html
• http//www-1.ibm.com/grid/about_grid/what_is.shtml
  
• Oinn, T.M. (2003) Talisman - rapid application
  development for the grid. Bioinformatics
  19(supp)i212-i214