TheGridBLOG

Grid Service Workflows:
Utilizing Service Data in Scientific Workflow Systems

abstract:
The emerging technologies of grid computing, web services, and service-oriented workflow will soon enable scientific projects to be conducted on a larger scale than ever before. Scientific workflows can be constructed by combining dispersed network accessible services into virtual organizations. Within a scientific workflow environment, metadata or grid service data is necessary for consumer application to discover services and for services to publish their properties.

This thesis proposes one of the first architectures and implementations of a grid service registry for use in scientific workflow applications. The registry uses the Globus Toolkit 3 and is made available as an OGSI compliant grid service. This thesis outlines several formats for the service data which allow consumer applications to attain features such as fault-tolerance, monitoring, and dynamic discovery or selection of services.

The Open Grid Service Architecture (OGSA) allows such registries to use service data for soft-state management, keeping track of metadata for service instances created from application factories. A service aggregation registry subscribes to services produced from a number of factories. Service data is express through an XSD namespace shared vocabulary. Discovery policy is expressed in XPath queries.

Grid Service Workflows:
Utilizing OGSA Service Data Notification Models for Service Composition

For this document, I will use the term 'grid service' to mean a Web Service which complies with the OGSI specification as well as the web services standards such as WSDL and SOAP. I will also assume that Grid Services are created with Java tools based on the Globus 3 Toolkit and OGSA, since no alternative for constructing OGSI compliant web services has been as thoroughly explored.

Several grid computing experts, including Ian Forester in The Physiology of the Grid, have expressed a desire for workflow services. Grid service workflows are called by a variety of names, including "aggregate services", "composite services", "meta services", "and control flow services." I present a definition based on commonalities in grid service and workflow literature. A "workflow grid service" is a grid service that takes a description of a workflow as input, and produces an adaptable interface to execute and monitor the workflow.


figure 1: a black boxed workflow service

I am concerned with what goes on inside the Workflow Service box to make such a service possible. I will not directly address the issue of dynamic user interfaces for workflows and grid services; the areas of Portals, APIs, and GUI controls are already crowded areas of research. Nor will I address the details of workflow description languages, web service composition languages, or grid services workflow languages. These areas are currently obscured by the politics of standards bodies and the voices of many researchers and corporations.

Though this research project of opening Workflow Grid Service black box, I will describe the current tools available for constructing such a service and identify the reasons why such services are not yet possible. I will also present several example partial service, and tutorial code for expanding this project once this field has matured.

Though this research I hope to enable future researcher to publish actual workflow grid services, which will in turn advance the construction of grid computing infrastructure in the US and abroad.

I'm at SC2003!

I attended the "How to Build a Grid Service Using Globus Toolkit 3" tutorial. It's super-cool.

The source files are here

The instructions for windows and linux

more info coming soon...

Let's take a closer look at the guys behind the "Old Wine" paper.

Wil M.P.van der Aalst has done a lot of work on workflow patterns. He's been doing workflow research for about 10 years and specializes in Petri nets.

Marlon Dumas is doing work on Web Service Composition and help create the Self-Serv P2P workflow system. He works with collaborates Boualem Benatallah and Quanzheng Sheng. He also works on electronic marketplaces and e-commerce trading agents just like NC State's own Peter Wurman.

Arthur ter Hofstede has also been doing a lot of work on workflows. He involved with the workflow group in CITI. He seems to have been doing work with BPEL, Workflow Paterns, and the Conference on Business Process Management.

Water, water everywhere, but not a drop to drink

Here's a message board thread form Alek illustrating the problem of too many workflow systems:
Too many workflows...


Here's a survey of workflow projects. I like this one.
Web Service Composition Languages: Old Wine in New Bottles?

Do you want to publish conference papers about Scientific Workflows? Who doesn't!

Here are some places to publish...

Collaborative Problem Solving Environments
and
Integrated Modeling of Distributed Systems & Workflow Applications


International Conference on Information Technology: Coding and Computing

Status Update

The scientific workflow survey of paper is stalled, while we finalize an outline/table of contents. I'm thinking something like this:

1. Intro and overview.
2. Workflow theory (patterns DAG, Petrie nets)
3. Workflow projects (short descriptions)
4. Comparison tables
5. Conclusion

I talked to Vouk today. He told me I should talk to our west-coast collaborators at SC2003. I'll be there from the 15th-19th.

These are my top ideas for a thesis topic:

* A comparison of XML-Based meta languages for grid service composition
* Upcoming conflicts and issues with WSDL 1.2 and BPEL4WS and GWSDL
* Techniques for dynamic endpoint swapping in grid service workflows
* A comparison of Grid Service Workflow Systems
* Software engineering patterns using web services and grid services