Quality of Service Ontology Languages for Web Services Discovery ...

Quality of Service Ontology Languages for Web Services Discovery: An Overview and Limitations

Furkh Zeshan1, Radziah Mohamad1, and Mohammad Nazir Ahmad2

1

Department of Software Engineering, Faculty of Computing, Universiti Teknologi Malaysia (UTM), 81310 Skudai, Johor, Malaysia [email protected], [email protected] 2 Department of Information Systems, Faculty of Computing, Universiti Teknologi Malaysia (UTM), 81310 Skudai, Johor, Malaysia [email protected]

Abstract. Web services discovery, ranking and selection based on QoS parameters is remained a hot topic for research since the start of the semantic web. Quality of service (QoS) plays an important role to resolve the issue of best service among the functional similar services. Semantic web relies on the ontologies for providing metadata schema and the vocabulary of concepts used in semantic annotation; resulting improved accuracy of web search. This is why; the success of semantic web depends on the proliferation of ontologies. Depending on the nature of the application, different companies may use different ontology languages and QoS models for web services selection which lead to the issue of heterogeneity. In this paper we have presented ontology evaluation criteria that if satisfied, can solve the problem of heterogeneity and interoperability. Moreover, ontology developers may also use these criteria to evaluate their developed ontology for the refinements. We have evaluated different ontologies in-order to know their strengths and limitations along with the new research directions.

Keywords: Web Services, Semantic Web, Semantic Web Services, QoS, Ontology.

1

Introduction

Service Oriented Computing relies on services for the development of dynamic, flexible, distributed and cost effective application. Service Oriented Architecture (SOA) is a style of design that guides at all steps of creating and using services throughout their lifecycle [1]. Service Oriented Architecture (SOA) enables different applications to exchange data and participate in different processes regardless the complexity of the applications. Semantic Web services with machine readable interface promises to enable the web services to be selected dynamically at run time to achieve the business goals within and across the organizational boundaries [3]. To find out the most relevant service among functionally similar that meet the requirements of users is the key issue in the web service discovery [4, 5]; however there is a need to define a set of well S. Yamamoto (Ed.): HIMI/HCII 2013, Part I, LNCS 8016, pp. 400–407, 2013. © Springer-Verlag Berlin Heidelberg 2013

Quality of Service Ontology Languages for Web Services Discovery

401

defined Quality of Services (QoS) criteria and user preferences [6]. Typically QoS resolves the issue of best service among the functional similar services by ranking and selection based on non-functional requirements. Hence QoS can be used as main factor for ranking the web services. Efficiently locating the web service registries and the efficiently retrieval of services with QoS from registries (UDDI) is one of the major challenge for web services [9]. This problem is becoming worst with the ever increasing of web services in the UDDI; the reason behind this problem is that the current web service technology is lack of semantic support. Researchers are trying to add semantics to web services to facilitate the web services discovery, ranking and selection process easy and call it as Semantic Web Services [9]. Ontologies are the proposed solution of heterogeneity among web services [7] and present the human knowledge as a critical component among the web services [8]. In this paper we have presented criteria that if satisfied, can solve the problem of heterogeneity and interoperability. We have evaluated different ontologies [15, 16, 18, 19, 20, 21, 22, 27] on these criteria to know, which one is the best ontology that can be used efficiently for web services discovery while addressing its limitations. The remaining paper is organized as; Section 2 presents state-of-the-art; Section 3 comparative evaluation, while Section 4 discusses the findings and finally Section 5 summarize the paper.

2

State-of-the-Art

In this section the review of different QoS ontologies have been presented. DAML-QoS [15]; It consists over three layers; profile layer, property definition layer and metrics layer. QoS property definition layer is used for elaborating the domain and range constraints of the property; QoS profile layer is used for QoS requirements matchmaking, while the metric layer is used for measuring the QoS properties. DAML-QoS is the enhanced form of DAML-S for QoS support. This ontology has a number of drawbacks; QoS ontology vocabulary is missing, concrete definition for QoS properties are also missing. For measuring the QoS information the scale and type of metric is not presented. The metric model is also not rich enough, approach is not perfect, and the use of OWL for imposing QoS values is also incorrect. QoSOnt [16]; ontology has three layers; base, attribute and the domain specific layers. Base layer consists on the generic concepts and the unit ontology. Attribute layer defines the particular metric of the QoS attribute. Dependability and the performance ontology are defined by a common set of QoS attributes introduced in [2]. The domain specific layer is used for defining the specific type of system, for example, web services or network system. This layer provides the QoS connection concepts in lower layer with OWL-S service profiles. QoSOnt is similar with [18, 19], but it does not support QoS profile and the QoS relationships. Conversion mechanism does not support units for mapping the different QoS properties; the usage aspect of the ontology is also poor. OWL-Q [18]; is an upper level ontology that extends OWL-S [14] specification for describing QoS of Web services. It consists of six facets; In Connecting Facet,

402 F. Zeshan, R. Mohamad, and M.N. Ahmad QoS attribute refers to service elements like preconditions, input, effects, parameters etc. It provides the high level concepts for defining QoS advertisements and demands. The Basic Facet is associated with several QoS offers or with only one QoS request. In this facet, QoS specification is used for the actual description of web services interms of validity period of the offer, cost of the service, associated currency with cost and security etc. QoS Metric Facet represents the upper ontology containing formal definitions of QoS metric model. It also represents the relationships between two metrics which may be independent and related. Function, Measurement Directive and Schedule Facets describe all the important properties and concepts for the definition of the metric function, while the Function and Schedule facets are used for computing and measuring the metrics. Unit Facet describes the unit of a QoS metric. A Unit is divided into BaseUnits, MultipleOfUnits and DerivedUnits but most of the time base units are used for measurements. Although OWL-Q ontology provides a detailed specification for modeling QoS information but it has weak support for the usage of the ontologies, for example QoS priority, quality level, QoS mandatory. It also does not include concrete definitions of common QoS properties. onQoS [19]; ontology was developed for specifying the QoS requirements by service consumers and by the service providers for advertising. It gives a clear description for a set of QoS properties. It is composed of three layers: upper, middle and lower. The onQoS upper ontology describes the QoS ontological language and answers the QoS information. Middle ontology refines the concepts of the upper ontology. Qos parameters, scales and metrics are defined in middle layer; this layer also gives the clear specification for metrics and the value types. In low level ontology domain specific properties, constraints and the QoS information are defined. But this ontology is quiet on specifying the QoS relationships, quality level, QoS priority, dynamism etc. QoS-MO [20]; has different quality levels for the requesters according to his demands, and can represent the interdependent QoS requirements between providers and requesters. This is an upper level ontology based on OWL-S and consists of three levels. First layer is for defining the general characteristics; QoS group related characteristics, for measurements and mapping the characteristics of the service. In second layer three types of constraints are defined, these constraints must be obeyed; service provider can specify the constraints for the service requester to follow in-order to make the service quality better. In third layer different QoS levels for different QoS usage models are defined. But this ontology has very weak support for metric transformation, units, value types and QoS priority. QoSHOnt [21]; author has presented QoS ontology for semantic service discovery, which provides a unified semantic model for users to advertising and querying the web services. Due to its context, service providers can describe different QoS specification for different circumstances. QoSHOnt consists of three parts, top level ontology is based on general requirements and it is independent to any application domain, middle level ontology is used to integrate quality features defined in the top ontology, while low level ontology is constructed for domain specific requirements. Author also has extended this ontology for contexts, by which service providers can describe different QoS specifications for different circumstances. The algorithm presented in this ontology has been used for matching and ranking the web services.


403

OSQS [22]; author has proposed ontology consisting of six components. OSQS is the extension of OWL-S for QoS. In this proposed ontology author has tried to integrate the quality standards like ISO/IEC 9126-1 and ISO/IEC 13236 [12, 13]. For web service discovery they also have extended the SPARQL query language for searching the web services according to the quality criteria but the quality properties are not used in query as other properties. Results of queries are produced without considering relationships that can be establish between quality properties. The extended SPARQL solve the problem to much extent. User discovers the services with required functionalities and then further classify on non-functional requirements. It does not provide the unit conversion facility as well as the metric is also not rich enough.

3

Comparative Evaluation

The defined criteria for ontologies evaluation are quite limited defined over the general requirements (questions defined above) and obviously not sufficient to fulfill all of the requirements of different applications. We have organized the criteria into three groups; Model related, Process related and runtime related criteria. We have compared different ontologies on these criteria, in-order to know the best approach that meets the maximum requirements. The short introduction of the criteria is given below. 3.1

Model Related Criteria

Short introduction of model related criteria is given below. Relationship: Relationship defines that how the qualities are related to each other [18]. Metric: Metric defines that how the values can be assigned to the QoS property [18]. Modularity: This requirement allows for adapting QoS ontology for different domains and applications. 3.2

Process Related Criteria

Short introduction of process related criteria is given below. Comparison: This property is an important factor for evaluating QoS metrics and their values. A QoS property can have one of five comparative effects: negative, positive, close, exact, and none.

Prioritization: In prioritization the requester defines that which qualities may be given more weight over others [24, 25].

Grouping: This feature allows for grouping several QoS properties which share similar characteristics or impacts in order to facilitate the evaluation and computing of the whole QoS ranking value of a Web service.

404 3.3

F. Zeshan, R. Mohamad, and M.N. Ahmad

Runtime Related Criteria

Short introduction of runtime related criteria is given below. Scalability: it defines whether the service capacity (ability to process more transactions in the defined period) can be increased as per requester requirements [10, 11, 23]. Dynamicity: Dynamicity defines the nature of the property. Dynamic properties are needed to be updated frequently while the static properties are updated once [24, 25, 26].

Interoperability: Means, that any two systems will drive the same inference from the same information, which is very difficult. It also deals with the compatibility of service with the standards [11, 23].

4

Discussion

In Table 1, we have analyzed different QoS ontologies to know their strengths and limitations. We have concluded that none of the existing ontologies cover all over the aspects of defined criteria; and still, there are many open issues for research. Most of the ontologies support the model related criteria, less of the process related and least in runtime related criteria. So, the runtime related aspects should be considered seriously for building an affective ontology. However, the phases (model and process) cover the maximum criteria aspects, but ontology support for runtime stage is still very small. Most of the ontologies described in the literature provide guideline for ontology development but still there is no agreement for best practices for ontology development. Table 1. Web service composition approaches

√

Interoperability

√

Scalability

Dynamicity

Grouping

√ √ √

Run-Time Related

√

Prioritization

√ √ √ √ √ √ √

√

Comparison

√

Modularity

[15] [16] [18] [19] [20] [21] [22] [27]

Process Related

DAML-QoS QoSOnt OWL-Q onQoS QoS-MO QoSHOnt OSQS OWL-SQ

Metric

Ontologies

Properties Relationship

Model Related

√


405

The results show (Table 1) that OWL-Q is the best ontology language among the rest as it fulfills the most of the parameters of the defined criteria. Despite of the recent efforts there is still a lot to do in enhancing the description of quality aspects of services. Some of our proposed research areas are as under.

Most of the ontologies described in the literature provide guideline for ontology development but still there is no agreement for best practices for ontology development. The number of best practices that must be followed by designers and developers is still very small, while there is no technical standard for ontology development. Fine-grained ontology development guidelines are needed to develop. In case of conflicting viewpoints fine-grained ontology can specify more accurately the vocabulary meanings.

Ontology support for runtime stage is very small. So, the runtime related aspects should be considered seriously for building an affective ontology.

5

Summary

Web services discovery, ranking and selection based on QoS parameters is remained a hot topic for research since the start of the semantic web. In this paper we have pointed out some problems associated with the current ontology models for effective discovery, ranking and selection of services based on QoS. In this paper we have presented criteria that if satisfied, can resolve these issues. We have evaluated different ontologies on these criteria to know, which one is the best ontology that can be used efficiently for web services discovery while addressing its limitations. We also have proposed the new research directions in this paper. Acknowledgement. We would like to thank Ministry of Higher Education Malaysia (MOHE) and Universiti Teknologi Malaysia (UTM) for sponsoring the research through the grant with vote number 4F165 and for providing the facilities and support for the research.

References 1. Newcomer, E., Lomow, G.: Understanding SOA with Web Services, ch. 1. Addison Wesley Professional (December 2004) 2. Laprie, J.C., Randell, B., Landwehr, C.: Basic Concepts and Taxonomy of Dependable and Secure Computing. IEEE Transactions on Dependable & Secure Computing 1(1), 11–33 (2004) 3. Cabral, L., Domingue, J., Motta, E., Payne, T.R., Hakimpour, F.: Approaches to semantic Web services: An overview and comparisons. In: Bussler, C.J., Davies, J., Fensel, D., Studer, R. (eds.) ESWS 2004. LNCS, vol. 3053, pp. 225–239. Springer, Heidelberg (2004) 4. Hu, J., Guo, C., Wang, H., Zou, P.: Quality driven Web services selection. In: Proceedings of the 2005 IEEE International Conference on e-Business Engineering, ICEBE 2005, Beijing, China, pp. 681–688 (2005)

406

F. Zeshan, R. Mohamad, and M.N. Ahmad

5. Seo, Y.J., Jeong, H.Y., Song, Y.J.: Best Web service selection based on the decision making between QoS criteria of service. In: Yang, L.T., Zhou, X.-s., Zhao, W., Wu, Z., Zhu, Y., Lin, M. (eds.) ICESS 2005. LNCS, vol. 3820, pp. 408–419. Springer, Heidelberg (2005) 6. Serhani, M.A., Dssouli, R., Hafid, A., Sahraoui, H.: A QoS broker based architecture for efficient web services selection. In: IEEE International Conference on Web Services, ICWS 2005, Florida, USA, pp. 113–120 (2005) 7. Chhabra, M., Lu, H.: Towards agent based web service. In: 6th IEEE/ACIS International Conference on Computer and Information Science (lCIS 2007), pp. 93–99 (July 2007) 8. Benatallah, B., Hacid, M.S., Leger, A., Rey, C., Toumani, F.: On automating Web services discovery, pp. 84–96. Springer (February 2004) 9. Wang, H., Zhang, Y., Sunderraman, R.: Extensible soft semantic web services agent. Soft Comput. 10, 1021–1029 (2005) 10. Giallonardo, E., Zimeo, E.: More Semantics inQoS Matching. In: Proc. of the IEEE Intl. Conf. on Service Oriented Computing and Applications, pp. 163–171. IEEE Computer Society, Los Alamitos (2007) 11. Yao, S.J., Chen, C.X., Dang, L.M., Liu, W.: Design of QoS ontology about dynamic web service selection. Computer Engineering and Design 29(6), 1500–1548 (2008) 12. ISO/IEC91 26-1, Quality characteristics and guidelines for their use. International Organization for Standardization I International Electrotechnical Commission. Tech. Rep. (2001) 13. ISOIIEC1 3236, Quality of service: Framework. International Organization for Standardization I International Electrotechnical Commission. Tech. Rep. (1999) 14. Ankolekar, A.: OWL-S: Semantic Markup for Web Services (2003), http://www.daml.org/services/owl-s/1.0/ 15. Zhou, C., Chia, L.T., Lee, B.S.: DAML-QoS Ontology for Web Services. In: Proceedings of the IEEE International Conference on Web Services, ICWS 2004 (2004) 16. Dobson, G., Lock, R., Sommerville, I.: QoSOnt: a QoS Ontology for Service-Centric Systems. In: Proceedings of the 2005 31st EUROMICRO Conference on Software Engineering and Advanced Applications (EUROMICRO-SEAA 2005) (2005) 17. Li, S., Zhou, J.: The WSMO-QoS Semantic Web Service Discovery Framework. IEEE (2009) 18. Kritikos, K., Plexousakis, D.: Semantic QoS Metric Matching. In: Proc. of European Conf. on Web Services, pp. 265–274. IEEE Computer (2006) 19. Giallonardo, E., Zimeo, E.: More Semantics in QoS Matching. In: Proc. of IEEE Intl. Conf. on Service Oriented Computing and Applications, pp. 163–171. IEEE Computer (2007) 20. Tondello, G.F., Siqueira, F.: The QoS-MO Ontology for Semantic QoS Modeling. In: SAC 2008, March 16-20 (2008) 21. Yu, C., Junyi, S., Yang, Y., Zhizhong, L.: A QoS Ontology for Semantic Service Discovery. In: 2009 International Conference on Networking and Digital Society (2009) 22. Jean, S., Losaviot, F., Matteot, A., Levy, N.: An Extension of OWL-S with Quality Standards. 978-1-4244-4840 2010 IEEE 23. Garcia, D.Z.G., Toledo, D., Felgar, M.B.: Semantic-enriched QoS Policies for Web Service Interactions. In: ACM International Conference Proceeding Series - Proceedings of the 12th Brazilian Symposium on Multimedia and the Web - WebMedia, vol. 192, pp. 35–44 (2006) 24. Papaioannou, I.V., Tsesmetzis, D.T., Roussaki, I.G., Anagnostou, M.E.: A QoSOntology Language for Web-Services. In: Proceedings of the 20th International Conference on Advanced Information Networking and Applications, pp. 101–106 (2006)


407

25. Tran, V.X.: WS-QoSOnto: A QoS Ontology for Web Services. In: IEEE International Symposium on Service-Oriented System Engineering, pp. 233–238 (2008) 26. Papaioannou, I.V., Tsesmetzis, D.T., Roussaki, I.G., Anagnostou, M.E.: A QoS Ontology Language for Web Services. In: Proc. of 20th Intl. Conf. on Advanced Information Networking and Applications, pp. 18–25. IEEE Computer (2006) 27. Guo, G., Yu, F., Chen, Z., Xie, D.: A Method for Semantic Web Service Selection Based on QoS Ontology. Journal of Computers 6(2) (February 2011)