CCIS 254 - A Comparative Evaluation of Web Service ... - Springer Link

A Comparative Evaluation of Web Service Discovery Approaches for Mobile Computing Nor Azizah Saadon and Radziah Mohamad Software Engineering Department, Faculty of Computer Science and Information Systems, Universiti Teknologi Malaysia, 81310 UTM, Skudai, Johor, Malaysia {azizahsaadon,radziahm}@utm.my

Abstract. Nowadays, with the advancement in mobile technologies, the use of web services has seen an explosion in interest for mobile computing environment. However, the heavyweight nature of conventional web services to be deployed on mobile devices brings new challenges in mobile computing in the coming future. Web services technology and mobile computing domains are converging at their intersection which leads to Mobile Web Services (MWS) which enables the service access in users’ mobile device. As the numbers of web services are increasing dramatically, the web service discovery becomes more important to discover the usable web services in effective and efficient manner. This paper discusses an overview of mobile challenges in MWS recent research, and a summary of the issues. Moreover, current approaches of Mobile Web Services Discovery are classified and these approaches are compared to some criteria. The results of this study will help researchers to deliver more applicable solutions with the most appropriate approaches for Mobile Web Services Discovery. Keywords: Web Services, Mobile Web Service, web service discovery, mobile computing, mobile web service discovery.

1

Introduction

Web services technology and mobile computing are converging at their intersection, known as Mobile Web Services (MWS). This concept has seen an explosion in interest and becoming more important in the development of web services in mobile computing. The worldwide revenue for MWS generated nearly USD 1.2 trillion in 2010 and is expected to break USD 1.7 trillion by the end of 2015 [1]. Nowadays, as the information is available anytime and anywhere from diverse mobile devices, the demands for MWS are growing dramatically as a new paradigm of web services in the mobile computing environment [11][13]. However, several challenges concerning the limitation of mobile devices such as the CPU processing power and smaller memory pose new and unique characteristics for web services in the mobile environment. A. Abd Manaf et al. (Eds.): ICIEIS 2011, Part IV, CCIS 254, pp. 238–252, 2011. c Springer-Verlag Berlin Heidelberg 2011

A Comparative Evaluation of WSD Approaches for Mobile Computing

239

MWS are web services in a mobile environment whereby a user has to search access and invoke services through a mobile device [5]. In MWS architecture, the mobile service providers publish WSDL (Web Service Description Language) descriptions of services to the mobile broker. Then, the mobile clients discover services from the mobile broker based on the WSDL descriptions. The characteristics of MWS architecture differ from conventional web services in terms of mobile devices which can be a client, provider, and even a broker. There are a lot of issues in MWS that need to be considered which are still in the early stage of research. Thus, the research questions that motivated this study are: – What are the issues of mobile web services that lead to the distinction of the unique characteristics compared to conventional one? – What are the mobile challenges which have driven to the various approaches in mobile web services? – What are the approaches that have been proposed for performance enhancements in mobile web services? Current conventional web services do not take into account about the specific characteristics for mobile computing environment. Therefore, it is necessary to consider the mobile challenges to fit into the web service technologies. This paper discusses an overview of different approaches in a mobile web services recent research and a summary of mobile requirements for web service in mobile computing. The aims of this study are at studying the concrete elements of MWS. The goal of this study differs from existing works, as this study aims to distinguish the unique characteristics of MWS compared to the conventional web services. The MWS issues are discussed and focused on the discovery of web service as it is an important part to discover the most relevant web services that match the request between clients’ request and service offered. Moreover, current approaches in MWS discovery are classified and these approaches are compared with some criteria. The results of this study will help researchers to deliver more applicable solutions with the most appropriate existing approaches to discover MWS. The layout of this paper is as follows: Section 2 discusses the mobile challenges in web service technologies and summarizes its mobile requirements. Section 3 discusses the MWS discovery issues as the focus of this work. Section 4 describes the classification of current approaches in MWS discovery. Section 5 describes the comparative evaluation with some criteria. Then, the discussion of comparative evaluation is given in Section 6. Finally, the conclusion of this work is given in Section 7.

2

Mobile Challenges in Web Services

Basically, MWS perform in a resource constraint environment compared to conventional web service architecture which is mainly designed for the environment of desktops and wired networks. To apply existing web services to mobile environment is not straightforward, since conventional web services do not meet

240

N.A. Saadon and R. Mohamad

the key aspects of mobile computing environment. Based on the exploratory literature study in MWS [2][4][5][13][14][16][17][20][21][23]the issues related to limitations of web services in mobile computing environment were identified. There are device capabilities, network connectivity, mobility, message exchange and security, in which each category has their own characteristics. The following are the details for the identified category. Device Capabilities: Pocket-sized mobile devices differ significantly from laptops or desktops in their shape, size, target user, functionality, features and service. Therefore, the physical constraints of mobile devices can be characterized to the ability of CPU processing power, smaller memory, smaller storage capacities, limited battery power, smaller screen size display, keypad or touch-screen capabilities and smaller bandwidth usage [11][13]. The limitation of processing power is tightly affected by the limitation of battery life because the battery life will decrease with a lot of computation. Besides that, the limited screen size display in mobile devices requires only important aspects of user interface to be displayed. The type of output display is another issue that regarding if the mobile devices can support the format. In addition, the limitation of keypad or touch-screen in mobile devices should consider the symbols and characters to be input and also the inputs amount should be kept in minimum. Work in [4], [5] and [21] consider the device capabilities in their approaches. The specific features of device specifications are exploited from the Composite Capabilities/ Preference Profile (CC/PP) standard. Network Connectivity: Limited bandwidth and high latency are the characteristics in MWS which normally will operate in wireless network connection[23]. Smaller bandwidth capability in mobile network leads to the limited data rates because the cost charged by mobile operator is expensive. Moreover, in wireless network environment, intermittent connectivity and network disconnection are the issues because of the stability of the wireless network itself. When the mobile user moves out of network range, this prompts to the disconnection. In such scenario, the request message or response message may fail to be delivered to the correspond requester. Work in [13] overcomes this issue by migrating the web services to adjacent mobile device. This is to provide continuous functionality with unstable network connectivity. Mobility: Mobility is one of the distinguished characteristics in MWS compared to the conventional web services. The portability and dynamicity of the mobile devices in mobile environment is the challenge where the changes of context and location information may change rapidly or unpredictably. Thus, work in [16] shows the examples that consider this issue. They consider the user context such as location properties as one of the elements in the discovery process. Message Exchange: Performance is the most significant issue in the message exchange for mobile environment due to its characteristics in mobile computing. To process XML (Extensible Markup Language) and SOAP (Simple Object Access Protocol) messages in MWS is quite challenging whereby XML messages


241

generate a larger overhead that is not suitable for mobile devices with limited processing power. Work in [2],[13]and[20] try to reduce SOAP messages exchange and to reduce size of messages by eliminating unnecessary message information in order to optimize the performance. Another approach in [14] proposes a getaway-based approach in message exchange and evaluates the response time performance. It is important to identify which part of the message is essential in MWS, which is still an open issue. Security: Security is an important aspect in mobile environment because mobile devices are portable and can be carried around. The IP address of the users device may change as the person moves to another network. Furthermore, with the limitation of network connectivity where communication can be disengaged, important applications such as mobile payment need to be considered thoroughly in the context of security if the network fails. Data and peer authentications are the important aspects in MWS and work in [17] considers the security aspect in their implementation. To summarize, the following mobile requirements as research challenges for the performance enhancement in MWS are identified: – Device specifications that can support available web services. – Heterogeneity and diversity of mobile devices with different capabilities and platform. – The changes of user context and web services context. – Unstable network connectivity requires migration of web services. – Requirement for lightweight message exchange for mobile device. – Security and authentications are important aspects in mobile computing. From the discussion above, MWS has to be designed within the limits of the available device constraints. Although the technologies of mobile devices are evolving, the development of MWS has to meet the mobile challenges. Therefore, in order to deploy web services in mobile computing environment, these mobile requirements should be taken into consideration.

3

Mobile Web Service Discovery

Web services have become one of the most promising approaches to establish service-oriented concept in mobile computing environment. Current research trends mostly focus on the entire web service life cycle, which are developing, deploying, publishing, discovering, composing, monitoring and optimizing the access to web services [27]. Nowadays, the overwhelming proliferation in web services which contributes to the similarity of web services functionality are increasing. This will cause the task to differentiate between several web services to become very complex and time consuming [5]. Therefore, finding relevant web services that match user request and what service provider offers is a challenge. Web service discovery has become essential to offer effective mechanism in order to discover the most relevant web services in mobile computing.

242


According to the W3C (World Wide Web Consortium), web service discovery is defined as ”an act of locating a machine-processable description of a Web service that may have been previously unknown and that meets certain functional criteria”[26]. The typical interactions during web service discovery process involve publishing, finding and binding process. Service providers publish and store their service descriptions in public service registry such as UDDI (Universal Description Discovery and Integration). Then, the service requestors or client will search the desired web service by looking up for the service descriptions stored in the public service registry. Candidates of appropriate web services are listed based on the user requirement. Service requestors then invoke the web service that matches their needs. This paper concentrates on the web service discovery for mobile computing and provides a classification of mobile web service discovery approaches in the next section.

4

Classification of Mobile Web Services Discovery Approaches

As discussed in the previous section, conventional web services do not meet the key aspects to be deployed in mobile devices. Therefore, research efforts in MWS have to be designed within the limits of the available device constraints. The existing discovery approaches based on their considered matching elements specifically in mobile computing environment are classified into three categories as in the following: 4.1

Context-Aware Discovery

Context information represents a part of non-functional properties in web services and can be interpreted differently from various perspectives. Different definition of context gives different impacts, specifically in mobile computing environment. For example, the context definition by Dey [9] is quiet general for this study in order to classify the web services approaches. This paper concentrates on mobile web service discovery and adapted Doulkeridis’s definition [10] as the basis to describe the context as ”the implicit information related both to the requesting user and service provider that can affect the usefulness of the returned results”. Context-aware discovery can be defined as the ability to make use of the context information to discover the most appropriate web services to the client against the service offered. Generally, two types of context are identified, which are the user context and web service context. User context can be the user profile, device capabilities, user preferences, location, temporal constraints and time, which are characterised by the users current situation. On the other hand, the properties for web service context can be the provider identity, location of the service, cost and payment method. All these context properties are considered in this paper to classify the approaches of MWS discovery to be in this group.


243

Among the approaches that consider the capability of mobile devices limitation is MobiEureka approach [4][5]. It considers device capabilities and user preferences by exploiting the Composite Capabilities/Preference Profile (CC/PP) standard defined by W3C. By using the information extracted from the device profile, relevant MWS are ranked based on the degree of relevance between the service offer and clients’ device. Another approach by Peng et al., [21] takes a step further in discovery mechanism by performing semantic matching between the service requester and service provider. It extends the service profile with the user profile information. A similar approach, namely AIDAS [25] also supports the semantic-based matchmaking by exploiting the user profile, device profile and service profile properties in the discovery process. AIDAS relies on its own ontology as it focuses on the service capabilities compared to the existing matchmaking algorithms and languages which are more focused on the service input and output. Approach in [13] proposes framework for hosting web services in mobile devices with the discovery manager as one of the components. Nonetheless, this approach is not focused specifically in discovery methods but still takes into account the device limitation. Work in [24] optimises the reasoning process to discover the most relevant web services in mobile devices. This approach presents semantic matching in mobile device by developing mTableaux algorithm but does not consider the ranking of web services or any methods of similarity calculation. 4.2

QoWS (Quality Of Web Service) -Aware Discovery

This category of matching elements concerns Quality of Service (QoS) aspects which are non-functional properties in web services. QoS refers to how well a service performs its behaviour to the customer. As the numbers of available web services are increasing, there will be a lot of competitors in service provisions that provide similar functionalities. Thus, when user requests for service, it will result in a long listing of available services in the users mobile devices. Nevertheless, with the constraints of mobile devices such as limited screen display, this will cause the user to scroll through the long list of web services candidates on mobile device. QoWS-aware discovery seems to be the most significant mobile service research recently, which can help the users to discover the most appropriate web services based on QoS properties. Alongside, Quality of Web Service (QoWS) is introduced in [27] to adapt the concept of QoS in web services. Generally, both QoS and QoWS refer to the similar aspects of quality as well. Work in [27] divides the QoWS into two types of quality: runtime quality and business quality. Runtime quality represents the response time, reliability, availability, accessibility and integrity. Business quality refers to the cost, reputation, and regulatory. This paper uses the quality parameters in [27] as a basis to classify the mobile web service approaches to be in this category. The approach in [3] introduces Web Service Relevancy Function (WsRF) for measuring the relevancy of requested service with QoWS properties. The input parameters from the clients are based on graphical user interface and then the

244


relevant services are ranked based on the WsRF. They adapted quality parameters such as response time, throughput, available, accessibility, interoperability analysis and cost. However, this study does not explain clearly on how the graphical user interface input method functions for this framework. Another approach in [6] proposes the Evolutionary Computing using Particle Swarm Optimization (PSO) technique in order to discover the most relevant services as QoWS properties as part of the query parameters. The considered QoWS are execution cost, execution time, availability, successful rate, reputation and frequency. This approach considers the semantic service discovery process plug-in together in composition phases by extending the semantic matching algorithm with simple PSO algorithm. QoWS score are calculated by similarity matching and relevant services are ranked based on QoWS score. Work in [28] proposes semantic matching by using vector space model to calculate semantic distance followed by an ontology hierarchical matching. Then, the concept of semantic metric matching for QoWS properties is introduced. Nevertheless, the authors of this paper did not mention the QoWS parameters that they considered. They developed upper ontology QoS-based web services that complement OWL-S but did not describe in details about the ontology on how QoWS properties will help the process of searching the relevant web services to suit the clients need. Another similar approach by Zhu and Meng [29] utilises OWL-S as a service description language and designs a prototype of service discovery which adopts Three-layer Pervasive Semantic Service Matching Algorithm (TPSSMA). The algorithm matches web services based on the category of service, input/output parameters and QoWS properties. Authors in this paper also did not provide the QoWS parameters that they considered. In addition, they did not explain how they will attach QoWS properties to the web service specifications. 4.3

Hybrid-CQ (Context-QoWS) Discovery

Another category that takes a step further, namely Hybrid-CQ (Context-QoWS) concerning both of the matching elements as discussed above: context and QoWS. However, taking both non-functional properties into consideration in service discovery is a challenging issue. There is a trade-off between the query response time and performance of relevant web services discovered as a lot of mechanisms are used to cater both properties. Work by Niazi and Mahmoud [19] considers the user preferences, device capabilities and also QoWS properties for discovering the most relevant MWS. Their approach is similar with MobiEureka approach [5] by using device capabilities and user preferences as the matching elements that exploit the Composite Capabilities/Preference Profile (CC/PP) specification. However, they enriched it by proposing ontology-based namely, Profile-Based Context-aware Ontology (PBCO). It is defined in OWL (Web Ontology Language) and it has been integrated with an ontology released by W3C recently, as


245

Delivery Context Ontology. QoWS properties that they considered are the availability, response time, throughput, latency and reliability which are attached in the PBCO. The advantage of this approach is the type of output format that can be displayed on mobile devices is considered. EASY approach in [16] introduces EASY-L (Language) that extends the language in OWL-S as the service capabilities for matching the functional and non-functional properties, and EASY-M (Matching) as for matching the service functional capabilities comparing two ontology concepts. Then, they calculate and rate the service between the matching service requests with the service advertisement. Their approach shows that combination of functional and nonfunctional properties (context and QoWS) provides the most relevant web services to the user and is even capable with the device constraint. ConQo[7] is another approach in this category that combines context and QoWS properties for the enhancement of web service discovery. ConQo utilises WSMO and introduces WSMO-QoS and WSMO-Context. Another approach in [12] proposes a personalised web service framework for discovering and selecting web services considering the context and QoWS. However, both papers, [7] and [12] did not expose the matching strategy and similarity calculation to provide ranking in details. They only provide general mechanisms to apply in service description and consider the subset of context and QoWS properties only.

5

Comparative Evaluation Criteria

This section discusses the comparison of above MWS discovery approaches with respect to some criteria. This work is mostly inspired by [8][15][18] and aims to consider, select and integrate the most appropriate criteria by those works for discovering MWS. In this paper, the criteria considered are: matching elements, matchmaking strategy, support similarity calculation, several stage discovery/matching and support QoWS. Matching Elements: The elements between the service requestor and service provider are considered for matching. The followings are the description of elements: – IO: Functional input and output for the web services. – PE: Preconditions and Effects of the web services. – Non-Functional Properties: User Context (device capabilities, user preferences, user profile, location, time, temporal constraints), Web Service Context (name, business category, provider identity, location of the service, cost, payment method) and Quality of Web Service (QoWS) – Other: Interface parts (operation names, operation numbers), WSDL descriptions and textual descriptions for the function of the service. Matchmaking Strategy: The matchmaking strategy uses the matching elements as the input or output of matching. The classified matching strategy below is based on the work in [22] and is categorised into four categories:

246


– Keyword-based: The query to find relevant web services based on a list of keywords as an input from client. Keyword-based matchmaking is easy to be implemented but it suffers a huge list of result as it cannot understand users actual intention. – Syntactic-based: The query uses keyword as a part of filter mechanism and then any algorithms are used to refine the list of web service candidates based on the web services function, interface, operation or structure. – Semantic-based: This strategy of matching exploits the advantage of ontology. The web service description file such as OWL-S, WSDL-S and WSMO are important in order to match the service request and service offer based on the ontology. – Pragmatic-based: This kind of strategy considers the context properties as a part of matching process. The matching strategy is any of the above strategy or combination of those three matching strategies. Support Similarity Calculation: The similarity calculation is based on the partially matching between the service requestor and service offer. With this partial match, the score to rank the appropriate services are calculated according to the desired input score. Therefore, the most relevant web services to the user are listed as the high score ranking. Any of the approaches that enable ranking of the matching results based on the user request is assigned to this criterion. Several Stages Discovery/Matching: Web service discovery with several stages leads to the meaningful and effective result. Any of discovery approaches that perform several stages of matching is assigned to this criterion. Accuracy: This criterion is defined as the accuracy of matching requirement during the discovery process between the service requestor and service offer. This criterion is used as a benchmark for the approaches that are being compared in this paper. The overall results of four criteria previously are used as the input score in the followings: – Matching Elements: The more elements supported for an approach, the more accurate it is. One score is attributed to it if any of the elements (IO, PE, Non-Functional Properties, Other) are supported by the approach. – Matchmaking strategy: The more matching strategy takes into account for an approach, the more accurate it is. One score is attributed to it if any of the strategies (keyword-based, syntactic-based, semantic-based, pragmaticbased) are considered by the approach. – Support similarity calculation: If an approach supports similarity calculation, it is more accurate compared to the approach without this support. Thus, one score is attributed to it. – Several stages discovery/matching: If an approach supports several stages matching, it is more accurate compared to the approach without this support. Thus, one score is attributed to it accordingly.


247

To determine the degree of Accuracy, it is evaluated as ’High’ if an approach in MWS discovery contributes a score between 8-10. It is evaluated as ’Average’ if the score is between 4-7 and set as ’Low’ if the score is between 1-3. Support QoWS Properties: Quality of Web Service (QoWS) refers to the non-functional properties that characterise web services overall behaviour. It becomes significant to measure the degree of desired web services with the combination of QoWS parameters. QoWS properties are adapted from the work in [27] as a basis of quality parameters in our work. It is divided into runtime quality and business quality [27]. – Runtime quality: It indicates the measurement of quality properties during the runtime of web services operation which are response time, reliability, availability, accessibility and integrity. If any of the approaches consider at least one of these parameters, it is assigned to this criterion. – Business quality: It represents the assessment of web service operation in the business perspective which are cost, reputation, and regulatory. If any of the approaches consider at least one of these parameters, it is assigned to this criterion. Quality: This criterion is defined as what reflects to the above Support QoWS properties. To determine the degree of Quality, if an approach in mobile web service discovery supports both of the above quality parameters: runtime quality and business quality, it is evaluated as ’High’. It is evaluated as ’Average’ if it supports either one of them and set as ’Low’ if none of the above parameters are considered.

6

Discussion

Table 1 shows the result of the comparison for MWS discovery approaches for each category with respect to the proposed classification of matching elements in MWS discovery. For this comparison, this study relies on the criteria in the matching elements, matchmaking strategy, support similarity calculation, support several stages discovery/matching, and support QoWS properties. The classified mobile web service discovery in the study is divided into three groups: Context-aware Discovery, QoWS-aware Discovery and Hybrid-CQ (Context-QoWS) Discovery. The classification of this study focuses on nonfunctional properties matching elements and the scope is restricted to web service discovery in mobile computing environment. The approaches that apply semantic-based matchmaking strategy rely on service inputs and service outputs (IO) of matching elements as those are the important parts in semantic solution to discover matching services [6][16][19][21][24] [25][28][29]. Service inputs and outputs are defined using ontology concepts semantically where OWL-S is used in these approaches as it is a widely used

248

N.A. Saadon and R. Mohamad Table 1. Classification of Mobile Web Services Discovery Approaches

language. However, the approach in [7] with semantic-based matchmaking strategy does not rely on the service inputs and outputs, as it considers non-functional aspects as the matching elements as well. All of the approaches compared were considered non-functional properties (NFP) as it is necessary in order to discover relevant web services instead of functional properties. Nevertheless, none of the approaches use preconditions and effects (PE) elements because of the optional condition which does not contribute much impact to the discovering process. The approaches in [12] and[19] use other matching elements whereas in [19], it considers interface, while in [12] it considers WSDL (Web Services Description Language) descriptions to express the input matching elements. For the matchmaking strategy, there are four categories that are keyword-based, syntactic-based, semantic-based and pragmatic-based. The discussion details for these groups of matchmaking strategy can be found in [22]. The approach in [12] uses keyword-based as a query. The disadvantage of using keyword-based is it cannot understand user’s real intention and it is a major drawback in matchmaking strategy. However, the work in [19] overcomes the capabilities of keywordbased as a filter mechanism and then using semantic-based and pragmatic-based matchmaking strategy in discovering process. Approaches in [3][5][13] use


249

syntactic-based whereas they use keyword as the first query process and then refining the result by adapting algorithm to discover the relevant web services. Approach in [5] uses device-aware computation algorithm and exploits metric to represent the device specification for discovering process. Similar approach in [3] represents the QoWS properties in metric and uses QoWS-aware computation algorithm to their approach. Semantic-based strategy is the most significant strategy that most of the approaches used [6][7][16][19][21][24][25][28][29]. It uses ontology to represent the matching elements semantically. Most of the approaches use pragmatic-based strategy except the approaches in QoWS-aware category [6],[28] and[29]. Pragmatic-based uses context as matching elements and it is a hybrid of any three matchmaking strategies. The context characterises current situation of the user’s context and web service context in order to make searching strategy efficiently match with the user request. Most of the approaches support similarity calculation except the approaches in [12],[13],[24]and[25]. Similarity calculation takes into account when partial match occurs. The relevant web services are ranked based on the calculated score. For example, the approach in [5] uses Device-aware Ranking Function (DaRF) to measure relevant mobile web services and ranks each of them according to the capability of device-specific features. This is similar to the approach in [3] which uses Web Service Relevancy Function (WsRF) to measure relevant mobile web services and ranks each of them according to the QoWS parameters that they have considered. For the several stages of discovery/matching criteria, most of the approaches consider these in their approach except work in [12],[13] and[25]. Web service discovery with several stages leads to the meaningful and effective result. However, there is a trade-off between the quality and response time of matching web services that leads to the complexity of discovery [18]. As the result, the accuracy criterion is the measurement for all the criteria discussed above. The approach in [19] resulted in a high accuracy in this comparison with most of the matchmaking requirements are fulfilled. Approaches in [3],[12],[13] and[25] are measured as having low accuracy as not many requirements are fulfilled. The remaining approaches, measured as having average accuracy with most of the approaches fall into this degree. For the Support QoWS criterion, it is divided into business quality and runtime quality. Approaches in category Context-aware Discovery obviously do not consider QoWS properties. Nevertheless, approaches in category QoWSaware Discovery mostly take into account the runtime quality but not all approaches in this category are considered having business quality. Approaches in [3],[6],[16],[21] and[28] are considered having business quality. Most of the approaches do not take into account the business quality because not all web services provide that information. However, it can be a benchmark with a lot of available web services recently to differentiate them. Quality is a criterion to measure the QoWS parameters as discussed above. Approaches in [3],[6] and[16] are measured as having high quality as they consider both of the QoWS parameters. All approaches in category Context-aware

250


Discovery are measured as having low quality as they do not consider both of the QoWS parameters. The remaining approaches are measured as having average quality as they consider either one of the QoWS parameters.

7

Conclusion

In this paper, the mobile requirements as boundary conditions and challenges for the development of web services in mobile computing environment are presented. The aim is to provide an overview of Mobile Web Services (MWS) and compare recent works in Mobile Web Services Discovery approaches. These approaches are based on non-functional properties matching elements and are classified into three categories: Context-aware Discovery, QoWS-aware Discovery and HybridCQ (Context-QoWS) Discovery. In each category, an introduction is given and the selected approaches are discussed. Then, the classified Mobile Web Services Discovery approaches with some criteria related to discovery aspects are compared. However, it cannot be claimed that this comparison is comprehensive and exhaustive. The considered criteria can be used as reference to help generally evaluating or selecting Mobile Web Services Discovery approaches. However, the result shows that there is no one prominent approaches fulfills all the criteria. Therefore, it depends on certain scenarios and specifications in order to decide the approach to be selected either in the development or for future research. Acknowledgments. We would like to thank Universiti Teknologi Malaysia for sponsoring the research through the grant with vote number 00J37 and for providing the facilities and support for the research. In addition, we would like to extend the gratitude to our lab members in EReTSEL Lab of Faculty of Computer Science and Information Systems, Universiti Teknologi Malaysia for their ideas and support throughout this study.

References 1. Portio Research Mobile Factbook (February 2011). Portio Research (2011) 2. Adacal, M., Bener, A.: Mobile Web Services: A New Agent-based Framework. IEEE Internet Computing 10(3), 58–65 (2006) 3. Al-Masri, E., Mahmoud, Q.H.: QoS-based Discovery and Ranking of Web Services. In: 2007 16th International Conference on Computer Communications and Networks, pp. 529–534. IEEE (August 2007) 4. Al-Masri, E., Mahmoud, Q.H.: Device-Aware Discovery and Ranking of Mobile Services. In: 2009 6th IEEE Consumer Communications and Networking Conference, pp. 1–5. IEEE (January 2009) 5. Al-Masri, E., Mahmoud, Q.H.: MobiEureka: An Approach for Enhancing the Discovery of Mobile Web Services. Personal and Ubiquitous Computing 14(7), 609–620 (2009) 6. Bhuvaneswari, A., Karpagam, G.: Reengineering Semantic Web Service Composition in a Mobile Environment. In: 2010 International Conference on Recent Trends in Information, Telecommunication and Computing, pp. 227–230. IEEE (March 2010)


251

7. Braun, I., Strunk, A., Stoyanova, G., Buder, B.: ConQo: A Context- And QoSAware Service Discovery. In: IADIS Intl. Conference WWW/Internet, pp. 432–436 (2008) 8. Cardoso, J.: Semantic Web Services: Theory, Tools, and Application. Idea Group Inc, IGI (2007) 9. Dey, A., Abowd, G., Salber, D.: A Conceptual Framework and a Toolkit for Supporting the Rapid Prototyping of Context-Aware Applications. Human-Computer Interaction 16(2), 97–166 (2001) 10. Doulkeridis, C.: CASD: Management of a Context-Aware Service Directory. Pervasive and Mobile Computing 4 (February 2004), 737–754 (2008) 11. Farley, P., Capp, M.: Mobile Web Services. BT Technology 23(3), 202–213 (2005) 12. Jarir, Z., Boumhamdi, K., Quafafou, M.: Context- and QoS Aware Personalization for Web Services: Web Information Extraction Case Study. In: 2009 Fifth International Conference on Signal Image Technology and Internet Based Systems, pp. 401–406. IEEE (November 2009) 13. Kim, Y., Lee, K.: A Lightweight Framework for Mobile Web Services. Computer Science-Research and Development 24(4), 199–209 (2009) 14. Mahmoud, Q.H., Ngo, T.: A Gateway-based Framework for Finding and Accessing Mobile Web Services. In: 2010 IEEE International Conference on Information Reuse & Integration, pp. 213–218. IEEE (August 2010) 15. Mohebbi, K., Ibrahim, S., Khezrian, M., Munusamy, K., Tabatabaei, S.: A Comparative Evaluation of Semantic Web Service Discovery Approaches. In: Proceedings of the 12th International Conference on Information Integration and Web-based Applications & Services, pp. 33–39. ACM (2010) 16. Mokhtar, S., Preuveneers, D., Georgantas, N., Issarny, V., Berbers, Y.: EASY: Efficient semAntic Service discoverY in Pervasive Computing Environments with QoS and Context Support. Journal of Systems and Software 81, 785–808 (2007) 17. Naumenko, A., Srirama, S., Terziyan, V., Jarke, M.: Semantic Authorization of Mobile Web Services. Journal of Theoretical and Applied Electronic Commerce Research 1(1), 1–15 (2006) 18. Ngane, L., Goh, A., Tru, C.: A Survey of Web Service Discovery Systems. International Journal of Information Technology and Web Engineering (IJITWE) 2(2), 65–80 (2009) 19. Niazi, R., Mahmoud, Q.H.: An Ontology-Based Framework for Discovering Mobile Services. In: 2009 Seventh Annual Communication Networks and Services Research Conference, pp. 178–184. IEEE (May 2009) 20. Oh, S., Aktas, M.: Efficiency Benefits in Mobile Web Service Architecture Using Context-store. Information Systems 3(2), 1–21 (2009) 21. Peng, R., Mi, Z., Wang, L.: An OWL-S Based Adaptive Service Discovery Algorithm for Mobile Users. In: 2008 4th International Conference on Wireless Communications, Networking and Mobile Computing, pp. 1–5. IEEE (October 2008) 22. Rong, W., Liu, K.: A Survey of Context Aware Web Service Discovery: From User’s Perspective. In: 2010 Fifth IEEE International Symposium on Service Oriented System Engineering, pp. 15–22. IEEE (June 2010) ˇ 23. Srirama, S., Vainikko, E., Sor, V., Jarke, M.: Scalable Mobile Web Services Mediation Framework. In: 2010 Fifth International Conference on Internet and Web Applications and Services, pp. 315–320. IEEE (2010) 24. Steller, L.A., Krishnaswamy, S., Gaber, M.M.: Enabling Scalable Semantic Reasoning for Mobile Services. International Journal on Semantic Web & Information Systems 5(2), 91–116 (2009)

252


25. Toninelli, A., Corradi, A., Montanari, R.: Semantic-based Discovery to Support Mobile Context-Aware Service Access. Computer Communications 31(5), 935–949 (2008) 26. W3C: Web service discovery, http://www.w3.org/TR/ws-arch/#wsdisc 27. Yu, Q., Liu, X., Bouguettaya, A., Medjahed, B.: Deploying and Managing Web Services: Issues, Solutions, and Directions. The VLDB Journal 17, 537–572 (2006) 28. Zhou, L.: An Approach of Semantic Web Service Discovery. In: 2010 International Conference on Communications and Mobile Computing, pp. 537–540. IEEE (April 2010) 29. Zhu, Y.A., Meng, X.H.: A Framework for Service Discovery in Pervasive Computing. In: Engineering and Computer Science (ICIECS), pp. 2–5 (2010)