2014 IEEE International Conference on Web Services
A Web Service Selection Framework based on User's Context and QoS Emna Khanfir
Chiraz El Hog
Raoudha Ben Djmeaa
Ikram Amous Ben Amor
MIRACL, ISIMS, Cité El Ons Route de Tunis Km 10 Sakiet Ezziet 3021, Sfax, Tunisia
[email protected]
MIRACL, ISIMS, Cité El Ons Route de Tunis Km 10 Sakiet Ezziet 3021, Sfax, Tunisia
[email protected]
MIRACL, ISIMS, Cité El Ons Route de Tunis Km 10 Sakiet Ezziet 3021, Sfax, Tunisia
[email protected] nu.tn
MIRACL, ISIMS, Cité El Ons Route de Tunis Km 10 Sakiet Ezziet 3021, Sfax, Tunisia
[email protected]
functionality, QoS and adaptation. The user’s context is automatically detected by a detection module. The rest of this paper is organized as follows: Section II discusses some related works. Section III describes the ontology of QoS. The QoS-ASF architecture is proposed on section V. Finally section VI concludes this paper.
Abstract—Since the emergence of Web services and the diversity of user's devices, an emerging need for adapted web services appeared. In fact, users aim to find web services meeting their requests, contexts and desired QoS. In order to achieve this goal, we propose in this paper a framework to find and select adaptable web services. Our framework is named QoS-ASF and it is based on a multi-agent system.
II. Keywords—AWS-WSDL; Web Service Selection; Adaptation; Quality of Service; Multi-agent System
I.
INTRODUCTION
With the emergence of new Web technologies, the numbers of web services have increased. The diversity of devices and connection used to access these web services have led to an adaptation. Ideally, when a user queries to get service that satisfies her request, he/she expect service adapted according to her/his context of use. According to Dey et al. [2], "context covers all the information that can be used to characterize the situation of an entity. The latter can be a person, place, or object relevant to the interaction between the user and the application, including the user and the application themselves". To render adaptable Web services, it becomes necessary to integrate context in the whole web service's life cycle. Although, Web service's standard architecture does not guarantee the selection of services taking into account the context of use. For this reason, a multitude of research focused adaptive selection mechanisms to solve this problem [1, 5 and 9]. In this paper, we propose a framework named QoS-ASF (QoS Adaptive Selection Framework) to discover and select adaptive web services based on context and QoS. A multiagent system is developed to carry out this framework. Our proposal is composed by two modules: a publishing module and a selecting module. The publishing module allows the provider to publish its adaptable web services and corresponding QoS. In this module, we adopted the description language AWS-WSDL proposed by ElHog et al [6] and we extend the AWS-Registry [7] as a registry. The selecting module allows the user to select the appropriate Web service that best satisfies his needs in terms of 978-1-4799-5054-6/14 $31.00 © 2014 IEEE DOI 10.1109/ICWS.2014.119
RELATED WORK
In this section, we take a look at some research works interested in the selection of adaptive web service. Elgazzar et al. [5] proposed a new Web service discovering method based on a clustering technique. This method consists of grouping services having similar functionalities into clusters in order to restrict the search space. Sriharee [10] described ontology-based selection architecture, in which service's rating are provided by reliable third-party organizations. Service features and rating description is allowed through OWL-S language. Then, the enriched service description will be used in matchmaking. Azmeh et al. [1] proposed a selection framework that enables users to get the suitable Web service. This framework takes into consideration user functional, QoS, and composition requirements and it is based on Formal Concept lattices. Lopez-Valisco et al. [9] proposed a selection framework that considerate the users need and context. This framework is an hybrid system that combines PUMAS [3] architecture with the SWA [8] architecture. Soukkarieh et al. [11] proposed new selection architecture of Web Services, supporting the adaptation to the user context. They combine the AHA architecture with the Web Service classical architecture. Then, they use an adaptation layer to deal with context information. All the previously studied works have a limited use of the notion of context and QoS. In fact, proposed selection methods consider the context or the QoS but not both of them at the same time. For this raison, we propose an adaptive selection framework taking together the context of the user and the QoS. Our framework aims to fulfill the user request in terms of QoS, preferences and context. III.
AWS-QREGISTRY
In this section, we give a QoS extension of the AWSRegistry proposed by Elhog et al [7]. This registry is an 708
extension of the UDDI registry to overcome its inability to support adaptive web service. Although the AWS-Registry gives answer to context publishing, it doesn't deal with QoS. To tackle this lack, we enrich the AWS-Registry data structure by two entities as shown in Figure 1
Figure 2. The ontology of quality of service OWL-QoS
x Reputation: is the degree of satisfaction which is given by users to evaluate the services. SR=(SR1,SR2,SRN)/N
with SR are the reputation service and N is the number of SR. x Scalability: the ability to increase the computational capacity of the system provider and its ability to handle more transactions or transactions in a given period.
Figure 1. The AWS-QRegistry data structure
The first entity is the QoS-registry; this entity contains the URL of the OWL-QoS ontology detailed in the section IV. The second entity Certificate-registry that represents the certificate of authenticity issued from the web service provider and contains the URL of the XML file authenticity. IV.
(1)
x Robustness / Flexibility: is the degree to which the service can operate correctly in the presence of invalid inputs, incomplete and contradictory.
ONTOLOGY OF QOS
x Capacity: the maximum number of simultaneous requests that the system supports with a great performance.
With an ever-increasing number of functionally similar web services, standard functional descriptions are becoming insufficient in the discovery process. Therefore, we need additional set of criteria to select the best service within many similar ones. In this paper we propose the quality of service as additional criteria. Some researchers have tried to add the QoS information during the service discovery process [12, 13, and 14]. But the syntactic descriptions of QoS are not adequate, since the service providers and requesters may use different concepts, scales and measurements. Hence, the semantics of QoS is necessary in the web service discovery step in order to render the suitable one answering the users’ needs. Therefore, we propose an extensible and scalable ontology named OWL-QoS that describes the semantics of QoS and defines the QoS attributes of web service as well as their relationships. OWL-QoS includes two parts of attributes. The first one defines generic attributes that are common of all services. The second part defines specific attributes for provider’s service domain. Figure 2 shown the detail of OWL-QoS attributes is given bellow: x Performance: the response time of web services.
x Security: ensures authentication, confidentiality and encryption of data from a web service. x Price: specify the cost implied by the invocation of the service. In the following section we propose an adaptable Web service selection framework based on QoS named QoS-ASF. V.
QOS-ASF ARCHITECTURE
The aim of our framework is to discover and select adaptive web services based on context and QoS. A multiagent system (MAS) is developed to carry out our proposal. We have chosen a multi-agent system (MAS) based on autonomous agents because our framework architecture requires a distribution of tasks between autonomous "entities" (or semi-autonomous) to achieve the publication and selection of web services and mainly to auto detect the user's context. As described in figure 3, QoS-ASF is decomposed by two modules. The first one describes the publishing module that allows the provider to publish adaptable web services and their corresponding QoS, in the AWS-QRegistry. The second module describes the selecting step. It allows the user to select the appropriate Web service that satisfies his needs in terms of functionality and adaptability. The user’s context is automatically detected by a detection module named W-
x Availability: the probability that the system is active. x Reliability: is the ability of service to perform its required function under given conditions for a specified period of time.
709
Registering agent extracts from the AWS-WSDL [6] the context and stores it in the context-register. Then, it forwards the WSDL description to AWS-Registry. It also stores OWL-QoS ontology in the QoS-Registry. The role of this agent is detailed in figure 4.
EDM (WildCad Extension Detected Module) that is an extension of WildCat module [4].
Figure 4. Adaptable Web service registration
B. Selecting Module In this section, we describe selecting module that allows user to select service that satisfies their need in term of context ‘user and QoS. This module is composed by three agent’s detailed bellow. Figure 3. QoS-ASF architecture
1) User agent: This agent retrieves request, preferences, QoS specified by user, and user's context that is automatically detected by a detection module named W-EDM (Wildcat – Extension Detection Module). The context’ user is composed by the characteristics of device, connection type and location. 2) Discovery agent: This agent enables the discovery of web services that satisfies functional and non-functional needs (context and QoS). This agent gets the query, the context information and the needed QoS from the User agent then it stores the list of discovered services into a parser. The service discovery process of this agent is enabled by four steps. In the first step, the Discovery agent accesses to the AWS-QRegistry in order to find services that satisfy the requirement of user. After retrieve a list of service L1 from AWS-QRegistry, discovery agent creates a first parser that contains signatures of web services found. In the second step, Discovery agent filters the previous step's results based on QoS attributes. To apply this refinement, Discovery agent uses algorithm 1. This algorithm browses the list L1 and matches each QoS attributes of a web service QosSWj. Attk to the respective QoS attributes desired by the user QosC.Atti. Then the discovery agent creates a new parser L2 that contains the list of Web service that meets QoS needs. In the third step, Discovery agent makes a filtering process based on the user's context. This process chooses from the list L2 selected in previous step, services that satisfies context information. This step is applied by the algorithm 2. This algorithm browses the list L2 and matches each service's context stored in the context Registry with the user's context retrieved using the W-EDM module. In the end of this algorithm, we obtain a set of web services having the ability to adapt their behavior to the collected context.
A. Publishing Module The publishing module is composed by three agents: provider agent, auditor agent and registering agent. These three agents are detailed bellow. 1) Provider agent: A provider agent is assign to each web service. This agent receives the URI of AWS-WSDL [6] and the OWLQoS ontology from the service provider. Then, it transmits the QoS ontology to the Auditor agent in order to verify the validity of QoS attributes. 2) Auditor agent: This agent verifies the validity of the QoS attributes defined by the provider in the OWL-QoS ontology. This step is required, because dishonest service providers could publish falsified QoS in order to increase their chances of being selected. So the Auditor agent verifies the QoS attributes and then sends a certificate to the Provider agent and stores one in the Certificate-registry. The validity of QoS is based in execution time and availability. To check availability of service, Auditor agent sends test message to web service for a period of time. Let D availability of service: D= uptime / (uptime + downtime)
(2)
where uptime is the total time when service is activate during measure period. Downtime is the total time when service is not activate during measure period 3) Registering agent: Before publishing the web service, this agent communicates with the Provider agent to verify the existence of a certificate. If the certificate exists, the
710
steps. The first one is services filtering based on QoS. The second step is to select services from first step result, which satisfy the context of user and his preferences. In this paper, we have also proposed an AWS-QRegistry that is an extension of AWS-Registry [7]. The AWSQRegistry enabled to publish the context of user and QoSOntology proposed that describes the quality of service criteria. In the future, we will integrate ranks classification to user's preferences in order to classify selected services by user's preferences. Also, we will extend our framework, in order to permit contextual composite context awareness services web and adapt the run of services web to the user’s context.
This set is the union of lists of web services resulting from each comparison of context attribute. LSWA = LSWALoc 䏖 LWSADivise 䏖 LWSAPref 䏖 ……
(3)
REFERENCES [1] [2]
[3]
[4]
[5]
[6]
[7]
In the fourth step, the agent Discovery classifies the list of step3, to order them by the rest of QoS criteria. So we calculate for all services the utility function: fj=6Vji*Pji
[8]
(4) [9]
With Vji value of a QoS attribute and Pji weight of a QoS attribute. Service having the higher value of utility function fj will be displayed at the top of the list of selected web services. 3) Selecting agent: This agent selects the web service which satisfies the user's request, context and has a better QoS. After, recovery the list of step 4, Selection agent skims this list and selects the service that has the highest fi function. The selected service will be displayed to the user through the User agent. VI.
[10]
[11]
[12] [13]
CONCLUSION
This paper proposed a framework named QoS-ASF to discover and select web services based on both user's context and QoS. Our framework is implemented through a multiagent system. The selection process is organized into two
[14]
711
Z. Azmeh, “A Web Service Selection Framework for an Assisted SOA.” Thesis, 2011. K.Dey and D.A. Abowd, “Conceptual Framework and a Toolkit for Supporting the Rapid Prototyping of Context-Aware Applications.” in Human-Computer Interaction, vol.16, pp. 97–166, 2001. M. P.Camps- Gleizes and V. P. Glize, “A theory of emergent computation based on cooperative self-organization for adaptive artificial systems.” In Proceedings of the Fourth European Congress of Systems Science (ECSS’99), pp., 1999. D.Pierre-Charles, and T. Ledoux, “WildCAT: a generic framework for context-aware applications.” in Proceedings of the 3rd international workshop on Middleware for pervasive and ad-hoc computing (MPAC '05), pp. 1-7, 2005. K. Elgazzar, A.E. Hassan, and M. Patrick, “Clustering WSDL Documents to Bootstrap the Discovery of Web Services.” In Proceeding of the IEEE International Conference on Web Services (ICWS’10), pp. 147-154, 2010. C.EL Hog, R. Ben Djemaa, and I. Amous. AWS-WSDL: a WSDL extension to support adaptive web service. In Proceeding of the 13th International Conference on Information Integration and Web-based Applications and Services (iiWAS’2011), pp. 477-480, 2011. C.El Hog , Ben Djemaa R., Amous I., Adaptable Web Service Registry for Publishing Profile Annotation Description. In Proceeding of the 10th IEEE International Conference on Ubiquitous Intelligence and Computing (UIC'13), pp. 533-538, 2013. C.Lopez-Velasco, Olivier, M.V. J. Gensel, and H.Martin, “ Adaptabilité a l’utilisateur dans le contexte des services web.” in MASIW’05 workshop, EGC’05 conference, pp. 153-158, 2005. C.Lopez-Velasco, A.Carrillo-Ramos, M.Villanova-Oliver, J.Gensel and H. Marti, “Sélection de services Web adaptés au contexte d'utilisation.” in Proceedings of INFORSID, pp.199-214, 2006. N. Sriharee, “Semantic Web Services Discovery Using Ontologybased Rating Model.” In ACM International Conference on Web Intelligence, 2006. B.Soukkarieh, F.Sedes, “Towards an Adaptive Web Information System Based on Web Services.” in Proceeding of International Conference on Autonomic and Autonomous Systems (ICAS’08), pp.272-277, 2008. S. P. Ran,.”A model for Web services discovery with QoS.” In ACM SIGecom Exchanges,pp.1-10 , 2003. V.X. Tran, H. Tsuji, R. Masuda,.”A new QoS ontology and its QoSbased ranking algorithm for Web services.” In Simulation Modelling Practice and Theory ,vol.17, pp 1378-1398,2009. S.J. Yao, C.X. Chen, L.M. Dang, W. Liu. “Design of QoS ontology about dynamic web service selection.” In Computer Engineering and Design vol.29, pp.1500–1548, 2008 .