Transcript tav-roadmap - IMPACT-MyGrid-Taverna

Taverna Roadmap
Shoaib Sufi
myGrid Project Manager
14/11/11
Sources of Influence
 Projects






SCAPE – Digital preservation
BioVel – Biodiversity science
CA4LS – Next Generation Sequencing
Wf4ever – Workflow preservation Astronomy & Bioinformatics focus
e-Lico – Data mining – (Kidney function data)
NeISS – Social Simulation
 Collaboration




IMPACT – Digital preservation
HELIO – Solar physics
Taverna - Galaxy – Bioinformatics
ViBRANT – Biodiversity
 User communities
 Super users requests (experts e.g. Andrea Wiggins)
 User requests
 Carole
 Feedback from conference attendance
 Developer community
 What are other projects doing
 Developer requests and contributions (integration)
 Useful to do / take into account
 Clouds
 Semantic Web
 Other promising or proven technologies
14/11/11
Taverna 2.3 Server Release 1
(November 2011)
 Hardened, debugged and tested
 Improved large input and output data handling
 Security support
 User authentication to the Server
 Secure service invocation from inside workflows
 Workflow run access control
 Improved job management and accounting trails
 Notification of workflow finishing using various mechanisms
14/11/11
Taverna 3
 Why – Standards compliance (OSGi), aid reuse
 Taverna 3 Platform OSGi beta and Command Line Tool
(January 2012)




Re-engineering the Taverna Engine using OSGi
Support programmatic access to the Taverna Core
New Run/Execution Manager and Edits API
Embeddable in applications for executing workflows
 Taverna 3 Platform OSGi and Command Line Tool (July
2012)
 Stabilized and hardened
 Plugin-enabled
 Taverna 3 Platform OSGi and Workbench alpha (July 2012)
 Based on the Taverna 3 Platform OSGi
 Taverna 3 Platform OSGi and Server beta (July 2012)
 Based on the Taverna 3 Platform OSGi
14/11/11
Draft Taverna 3 features
 Better user-interaction during workflow run
 - Improved interaction with visualizing tools and services
 ‘Blackboard’ support
 Improved workflow diagram
 Mark as boring
 More obvious looping & iteration strategy
 Data
 JSON related services (create, search, extract) (in 2.x also)
 XML related service (create via templates) (in 2.x also)
 Table support (structured objects)
 Services
 Choice of languages for in-workflow-programming via Java Scripting
framework
 A better ‘bean shell’ (e.g. python or other languages supported by JVM)
 Components (semantically annotated workflows as pipeline steps)
 - Components creation/deletion
 - Components registry
 Enhanced execution
 Ability to change WSDL bindings
 Intelligent re-run/resume of workflows (use results before error occurred)
 Tarpit mitigation – i.e. XSD caching (2.x also)
14/11/11
Taverna 3 features continued …
 Improved Provenance
 More efficient provenance capture
 More detailed provenance capture
 Export of Provenance in OPM and W3 PROV serialisation
 Research Object support
 via SCUFL 2 Data bundle:
 Workflow
 Provenance
 Data
 In
 Out
 Intermediate
 Manifest
 Control flow
 Branching support (if/else)
 Better looping (explicit while)
14/11/11
Taverna 3 features continued(2) …
 Domain specific support (from projects req’s)
 Astronomy




VOTable visualisation support (Workbench)
VOTable local workers support (Engine)
Opens way for Table support +XML template use case
SAMP support (Workbench) local blackboard
 Biodiversity
 Asynchronous input support
 Sensor data
 Relevant to other instruments producing continuous data
 Security
 Single Sign On
 Server
 Pause/resume
 Fine grained monitoring and feedback for a running workflow.
 Where is it upto
14/11/11
SCUFL2 – the new workflow format
 SCUFL2 Language Specification (January 2012)
 Taverna workflow language specification more like SCUFL from
Taverna 1.7.x to replace the current t2flow serialisation format
 Easier for third party use
 A packaged format with annotation support
 SCUFL2 Tools beta (January 2012)
 Tools for conversion from t2flow to SCUFL2
 SCUFL2 Tools Release (July 2012)
 A stabilised and fully tested version
 Plugin for the Taverna Workbench 2.3
 Plugin for the Taverna Server 2.3
 Other SCULF2 tooling
 workflows to Hadoop/MapReduce (SCAPE project)
 Workflows to Wf4ever abstract workflow format
 Processers and links (no activities) – Skeleton workflow
 Conversion from abstract format to SCULF2 (won’t execute but act as
template)
14/11/11
Cloudy with a chance of Taverna
 Taverna Infrastructure Virtual Appliance
 Taverna Server Virtual Machine (VM) (December 2011)
 Including Taverna 2.3 Server and management interface
 Amazon Machine Interface (AMI) (December 2011)
 Including Taverna 2.3 Server and management interface
14/11/11
KISS – Taverna keeping it simple
 (December 2011 alpha)
 Web based interface
 Connecting workflows
 Hides complex iterations/looping/config in ‘components’
 Build ‘component’ workflows in Taverna Workbench and then
deploy in KISS
 Simple install and one config file containing workflow
definitions for a particular domain/instance
 Uses Wireit
 Will shortly have a Baclava data viewer
 Alpha (uses command line)
 Multi user support Q1 2012
 Will use Taverna Server
14/11/11
14/11/11
Plugins
 Data Mining plugin (December 2011)
 Makes all the RapidMiner (www.rapidminer.com) services
available in Taverna
 more than 500 operators for data integration and transformation, data
mining, evaluation, and visualization
 Automatic workflow building via metadata and goals
 SPARQL plugin (2.x and 3)
 Query RDF data sources (e.g Linked Data)
14/11/11