Seite - 358 - in Intelligent Environments 2019 - Workshop Proceedings of the 15th International Conference on Intelligent Environments

iii) Service Requester (sometime called service client) which searches services in the service registry and consumes themby sending requests to the service providers.Web services are based on standard technologies and protocols (based onXML) that allow applications to communicate remotely via the Internet, regardless of the platforms and languages onwhich they are based. They arewidely used by companies, which allow themtoexposeanumberof servicesandexchange informationbetween them. Due to the great success ofweb services the number of deployedweb services in thewebisgrowingfast.Asimplerequest fromaclientcanhavehundreds(or thousands) of similar services as result.As consequence, the client has a serious issue for selecting the “best” service among the returned list of services.Wemean by “best” service, the service that responds to the client requirements. In fact, the Quality of Service (QoS) is considered oneof themost importantmetrics for distinguishingbetween services. In general, the returned services are not sorted based onQoS, but they are sorted based on their registration date in TheUniversalDescriptionDiscovery and Integration stan- dard(UDDI) registry[3]. In the literature, several solutions have been proposed to deal with the problemof selecting the best services. Someof these solutions are basedon i) syntax/semanticbasedapproaches[4][6], ii)userprofilebasedapproaches[7], iii)genetic algorithmbasedapproach[8] iv)QoSbasedapproaches[5].Generally, theQoSbasedap- proaches aremost popular in the industry. TheQoS can be defined as the description ormeasurementof theoverall performanceof a serviceby taking into consideration the non-functional properties such as: reliability, availability and response time[9]. Aweb service consists of a functional part that describeswhat the servicemust do and a non- functionalpart thatdescribeshowtheservicecando it. Besides, selecting services based on QoS is like solving amulti-criteria problem by searching a service that meets several non functional properties. Multiple Criteria DecisionMaking (MCDM) is a family ofmulti-criteria decision analysismethods.We can cite for example, Elimination and Choice Translating algorithm (ELECTRE)[10], Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS)[11] and AHPdevelopedbySaaty[12],...etc. In this paper,we propose an approach based on theMCDMmethods for selecting thebest services.Ourapproach is composedof twomainsteps:First,weassignweights to eachnon-functional property that qualify aWebservice.Weuse in this step theAHP method.Second,weuse theseweights to rank the list of the similar services inorder to select thebestone.For rankingservicesweuse theTOPSISmethod. Inorder toevaluateourapproachweshowits applicabilityonanexampleof select- ing thebest serviceamongasetof real services related toSMSsendingwebservices. This paper is organized as follows: Section 2 presents some relatedwork. Section 3describesourproposedapproach. In theSections4and 5wepresent howweapplied theAHPandTOPSISmethods.Before concludingwepresent inSection6 theobtained experimentation results.Finally,Section7presentsaconclusionandfuturedirectionsof thiswork. 2. RelatedWork In[15], the authors proposed aweb serviceQoSManagermodule responsible formea- suringQoS information for the collectedWeb services, and information is stored in the WebServiceStorage. H.Belouaaretal. /AHPandTOPSISMethods forWebServiceSelection358