1. Service Composition in the Context of Grid Xiaofeng Du (Presenter), William Song, Malcolm Munro University of Durham {xiaofeng.du, w.w.song, malcolm.munro}@durham.ac.uk AHM2006, 21.09.2006

2. Outline Problems and issues. Web Service and Grid Service Conceptual Service Composition Semantic Service Description Aspects and Matchmaking Summary and future work

3. Introduction How to extract and represent semantics for the Web information and the Grid recourses has became a consensus issue in WWW and Grid research. Grid provides an infrastructure for resource sharing and cooperation. Interoperability and collaboration requires common understanding of information/data and resources. The understanding requires canonical and well- formed semantic description.

4. Web Services and Grid Service A Web service is a software system designed to support interoperable machine-to-machine interaction over a network that has an interface described in a machine- processable format. A Grid service is a Web service that conforms to a set of conventions (interfaces and behaviours) that define how a client interacts with a Grid service. The conventions adhere to the OGSI specifications. The context of the Grid is e-Science. Grid is the underlying computer infrastructure that provides these facilities to facilitate e-Science. Web Service, realized in Grid services, uses a standardized XML messaging system and a stack of communication protocols to makes software available over the Web.

5. Type of Services (1) Services Types: To exchange information and documents; To interactively and cooperatively perform functions; To share available resources.

6. Type of Services (2) The first type is the Web systems or software for information exchange. The second one, we consider that web service is a chain of (sub-) services in order to accomplish a user given task, see the diagram. Grid service is considered to be of the third type, which emphasizes cooperation and sharing of resources (including data) among entities.

7. Task, Service, Resource, and Grid The Relationships among task, web service, resource, and grid infrastructure are illustrated in the diagram. In the grid infrastructure, the resource sharing is task driven. A resource consumer issues tasks to request resource sharing from the resource providers. The web services can be considered as an interface between tasks and resources, which can consume suitable resources to achieve a certain task.

8. Service Composition (1) Motivation Build more powerful service using the basic existing services. Better fulfil service requesters requirement. Enhance resource reuse and reduce the cost and time of new service development. Service composition process Identifying sub tasks, Locating suitable sub-services, Formatting the sub-services into a service flow, Executing the service flow to achieve a task which is the goal of the composite service.

9. Service Composition (2) Data type and Semantics need to be considered when composing services. Some other contextual factors are also need to be considered, such as: Time – multiple services invocation and synchronisation. Data Consistency – data type, semantics, and units etc. Location – services location. etc.

10. Service Composition (3) Inputs and outputs of a composite service are generated from some of its internal sub- services inputs and outputs. Use directed graph G(V, E) to represent all the services in the grid environment Use Sub_G 1 (V 1, E 1 ) to represent a composite service Use Sub_G 2 (V 2, E 2 ) to represent a composite service together with its context. Then context of the composite service can be represented as graph C(V c, E c ) Then the set of arcs E c represent the inputs and outputs of the composite service. v1v1 v1v1 v1v1 v1v1 v c vcvc Sub_G 1 C vcvc v v v v v v v Sub_G 2 v … … The Grid (G) Note: the arcs in the graph G are uncertain due to different composition requirements. Here just captures a possible situation.

11. More on Composition Composition is important in services and resources reuse. Manual work is unsatisfactory. Automatic composition is preferred. A comprehensive description model or framework is crucial to achieve automation.

12. Semantic Service Description Aspects IOPE: addresses the services input data type, output data type, pre- condition, and effects. It also addresses the semantic meaning of the inputs and outputs. Metadata: addresses the non-functional description of the service including identifier, natural language description, location, quality attributes, and category. Ontology: addresses the concept and domain of the service. Upper Compositionality: addresses which kind of services can be composed by using this service. This aspect indicates the relationships between this service and other services. Lower Compositionality: addresses what kind of services can be used to compose this service. This aspect also indicates the relationships between this service and other services. Resource: addresses what kind of resources the service will consume. The resources include renewable resources, such as bandwidth, memory allocation, and CPU usage, and consumable resources, such as time.

13. Contextual based Semantic Model The notation used in the diagram: S i =Service i Service can be either the parent or ancestor of Service i S 1, S 2, S 3, and S 4 are any services. I, P is the Inputs and Pre- condition, and O, E is the Outputs and Effects. is_a_component_of Relationships among semantic description aspects.

14. Matchmaking Semantic Distance Calculation. Based on the model introduced previously, a semantic distance calculation formula is proposed. Here λ is the set of all the semantic characteristics functions. The result value will be between 0 and 1. This formula is not a new matchmaking method, but to examine how the model can benefit the service matching.

15. Matchmaking Example Service Requirements Number of inputs: 2, Input data type: double, Output data type: double, Description keyword: addition, A mathematics calculation service. Can be used to calculate the perimeter of a rectangle Matching AspectsService1 double addition(double a, double b) Service2 int addition(int a, int b) Service3 double power(double a, double b) IOPE1.00.51.0 Ontology1.0 Metadata0.667 0.0 Upper Compositionality1.0 0 Sum3.6673.1672.0 Distance 0.920.790.5 The Chosen service is Service1

16. Summary and Future Work Defining a contextual based semantic model to describe services is important in semantic web service, grid computing, and web services researches. It is particularly significant in semantic based automatic service searching, matchmaking, and composition. Consummate the contextual based semantic model to more sufficiently describe services. Develop an advanced quantitative analysis algorithm for semantic matching and to evaluate its performance.

17. References 1.Foster, I. and Kesselman, C., The Grid: Blueprint for a New Computing Infrastructure, Publisher: Morgan Kaufmann Inc, 1999. ISBN: 1-555860-475-8 2.Foster, I., Kesselman, C., and Tuecke, S., The Anatomy of the Grid: Enabling Scalable Virtual Organizations. International Journal of High Performance Computing Applications, 15 (3). 200-222. 2001 3.Dey, A. K. and Abowd, G. D., Towards a Better Understanding of Context and Context Awareness. Presented at the CHI 2000 Workshop on the What, Who, Where, When, Why and How of Context-Awareness, April 1-6, 2000 4.Fensel, D. and Bussler, C., The Web Service Modeling Framework WSMF. Electronic Commerce Research and Applications, Vol. 1, Issue 2, Elsevier Science B.V. 5.Lara, R., Lausen, H., Arroyo, S., Bruijn, J., and Fensel, D., Semantic Web Services: Description Requirements and Current Technologies, In International Workshop on Electronic Commerce, Agents, and Semantic Web Services, September 2003. 6.Medjahed, B., Bouguettaya, A., and Elmagarmid, A. K., Composing Web Services on the Semantic Web, the VLDB Journal 2003. Volume 12, pp. 333- 351 7.De Roure, D., Jennings, N. R., and Shadbolt, N. R. The Semantic Grid: Past, Present and Future. Procedings of the IEEE, 2005, Volume 93, Issue 3, pp. 669-681

18. Thank you for your attention and patience ! Any Questions?