A BROKER-BASED WEB SERVICE ARCHITECTURE FOR SERVICE REGISTRATION AND DISCOVERY WITH QOS

A BROKER-BASED WEB SERVICE ARCHITECTURE FOR SERVICE REGISTRATION AND DISCOVERY WITH QOS Dr. T.Rajendran Professor and Dean, Department of CSE &IT, Angel College of Engineering and Technology, Tirupur, India. Email: rajendran_tm@yahoo.com Abstract The increasing number of Web services over the web makes the consumer to apply tools to search for appropriate Web services available though out the world. Web service technology provides a way for simplifying inter-operability among different platforms. There exist many Web services which reveal similar functional characteristics; non-functional properties (QoS) have become essential criteria to enhance the discovery and registration process of services. The purpose of web service discovery is to choose best possible web service for a particular task. When dynamic discovery is used in Web Services, it is common that the result of the discovery contains more than one web service provider. In this paper, we proposed a novel approach for designing and developing a broker-based architecture (WSB) and its QoS-based matching, ranking and selection algorithm for evaluating and discovering the best possible Web services for a specific task. We explained the underlying ideas of our approach and experimental results have shown the effectiveness of our approach. Keywords--- Broker-Based, Web Service Architecture, Service Registration, Service Discovery, Quality of Services (QoS). I. INTRODUCTION Service-Oriented Architecture (SOA) is an architectural approach for constructing complex software-intensive systems from a set of universally interconnected and interdependent building blocks, principles used during the phases of systems development and integration. A deployed SOA-based architecture will provide a loosely-integrated suite of services that can be used within multiple business domains. A SOA is essentially a collection of services. These services communicate with each other. The communication can involve either simple data passing or it could involve two or more services coordinating some activity. Some means of connecting services to each other is needed. Web service is a software system designed to support interoperable machine-to-machine interaction over a network. Web Services interact with each other, fulfilling tasks and requests that, in turn, carry out parts of complex transactions or workflows. If multiple Web services provide the same functionality, then Quality of Service (QoS) requirements can be used as a secondary criterion for service selection. QoS is a set of non-functional attributes like service response time, throughput, reliability, and availability [1], [2]. The current Universal Description, Discovery and Integration (UDDI) registries only support Web services discovery based on the functional aspects of services [1]. The problem, for that reason, is firstly to accommodate the QoS information in the UDDI, and secondly to guarantee some extent of authenticity of the published QoS information. QoS information published by the service providers may not always be accurate and up-todate. There are two major problems in using QoS for Web service discovery. First is the specification and storage of the QoS information, and second is the specification of the customer s requirements and matching these against the information available. Major efforts in this area include Web Services Level Agreements (WSLA) [3] by IBM, Web Services Policy Framework (WS Policy) [4], and the Ontology Web Language for Services (OWL-S) [5]. Most of these efforts represent a complex framework focusing not only on QoS specifications, but on a more complete set of aspects relating to Web services. Some researchers propose other simpler models and approaches [6]-[8] for dynamic Web services discovery. However, they all struggle with the same challenges related to QoS publishing and matching. Nowadays, both Web services providers and clients are concerned with the QoS guaranteed by Web services. From the client point of view, Web service based QoS discovery is a multi-criteria decision mechanism that requires knowledge about the service and its QoS description. However, most of the clients are not experienced enough to acquire the best selection of Web service based on its described QoS characteristics. Sri Sai Ram Engineering College, An ISO 91:28 Certified & NBA Accredited Engineering Institute, Chennai, INDIA. Page 1

They simply trust the QoS information published by the provider; however, most of the Web services providers do not guarantee and provide assurance to the level of QoS offered by their Web services. Based on the above, we propose a Web services discovery architecture that contains an extended UDDI to accommodate the QoS information, and Web Service Broker (WSB) to facilitate the service discovery. We develop a service matching, ranking and selection algorithm based on a matching algorithm proposed by Maximilien and Singh [9]. Our algorithm finds a set of services that match the consumer s requirements, ranks these services using their QoS information and feedback rating, and finally returns the top M services (M indicates the maximum number of services to be returned) based on the consumer s preferences in the service discovery request. QoS delivered to a client may be affected by many factors, including the performance of the Web service itself, the hosting platform and the underlying network. A set of verification procedures is essential for providers to remain competitive and for clients to make the right selection and trust the published QoS metrics. For the success of any QoS based Web services architecture, it should support a set of features: 1) QoS Verification and Certification to guide Web services discovery and 2) QoS aware Web services publishing and discovery. In this paper, we propose a broker-based architecture for Web service registration and discovery with QoS characteristics. The role of the Web Service Broker (WSB) is to support QoS provisioning and assurance in delivering Web services. It implements the concept of Quality Analysis, and QoS verifying and certifying process. The goal of this research is to investigate how dynamic Web service discovery can be realized to satisfy a customer s QoS requirements using a new architecture that can be accommodated within the existing basic Web service protocols. The rest of the paper is organized as follows. Section 2 outlines the related research conducted in the area of Web Services Discovery, QoS and reputation. In Section 3, the proposed Web Service Broker architecture for Web service registration and discovery is analysed and discussed. Section 4 presents experimental results which evaluate the effectiveness of our architecture. Section 5 concludes the paper and presents possible future research work. II. RELATED WORK Researchers have proposed various approaches for dynamic Web service discovery. Maximilien and Singh [9] proposed a multi-agent based architecture to select the best service according to the consumers preferences. Blum [1] proposes to extend the use of Technical models (tmodels), within the UDDI to represent different categories of information such as version and QoS information. Ran [1] proposes an extended service discovery model containing the traditional components: service provider, service consumer, and UDDI registry, along with a new component called a Certifier. The Certifier verifies the QoS of a Web service before its registration. The consumer can also verify the advertised QoS with the Certifier before binding to a Web service. Although this model incorporates QoS into the UDDI, it does not integrate consumer feedback into the service discovery process. However, it lacks support for the dynamism of Web services. Crasso et al. [31] discussed eight shortcomings in the present Web service description and have termed them as anti-patterns which prevent the efficient discovery of the services. They have experimentally shown that discovery is more accurate if all such anti-patterns are removed. Different domains and applications may require different QoS properties; therefore we need a more efficient and flexible method to express QoS information [32]. Mydhili and Gopalakrishna [33] had given an exhaustive review summarizing the results, observations, and findings of the renowned researchers in the domain of Web Service Discovery. They highlighted the list of problems, which need to be looked into and investigated. The comprehensive listing of the open-ended unresolved issues was presented in a novel way, by providing its Cause-Effect Analysis. Shrabani Mallick and Kushwaha [11] proposed Web service discovery architecture based on x-soa, organizes the method of Web service discovery in an efficient and structured manner using an intermediate, Request Analyser (RA), between service provider and service consumer. Algorithm in Shrabani Mallick and Kushwaha [11], LWSDM, works for a complete cycle of Web service discovery. Majithia et al. [12] proposed a framework for reputation-based semantic service discovery. Ratings of services in different contexts are collected from service consumers by a reputation management system. T. Rajendran and P. Balasubramanie [13] showed a framework for agent-based Web services discovery with QoS to select the suitable Web service that satisfies the clients preferences and QoS constraints. It contains an extended UDDI to accommodate the QoS information. IBM proposes Web Service Level Agreements (WSLA), which is an XML specification of SLAs for Web services focusing on QoS Constraints [14]. Many of these approaches do not provide guarantees as to the accuracy of the QoS values over time or having up-to-date QoS information. Sri Sai Ram Engineering College, An ISO 91:28 Certified & NBA Accredited Engineering Institute, Chennai, INDIA. Page 2

Farkas and Charaf [15] proposed software architecture to provide QoS-enabled Web services by adding a QoS broker between clients and service providers to discover the QoS aware services in UDDI. However, no detailed and accurate information about QoS broker, such as how it is designed and the functionality of it is presented so far. UDDI extension to support QoS-enriched service publication and discovery has generated several research efforts. Shaikh Ali s approach [16] was based on the extension of the UDDI business service structure, but potential QoS changes are not considered. Chen et al. [17] proposed a registry that receives reports made by consumers to generate QoS summaries for invoked Web services. Kalepu et al. [18] evaluated the reputation of a service as a function of three factors: ratings made by users, service quality compliance, and the changes of service quality confirmance over time. However, these solutions do not take into account the trustworthiness of QoS reports produced by users, which is important to assure the accuracy of the QoS-based Web service selection and ranking results. Liu et al. [19] proposed an approach for rates services computationally in terms of their quality performance from QoS information provided by monitoring services and users. The authors also employ a simple approach of reputation management by identifying every requester to avoid report flooding. In Diego Zuquim Guimaraes Garcia and Maria Beatriz Felgar de Toledo [2], an extended Web service architecture to support QoS management was explained. The architecture is currently being integrated with Business Process Management (BPM) Technology. The major contributions are: Extending the WS policy framework to specify QoS policies for Web services, extending the UDDI information model and API set to refine service discovery and using tmodels to define QoS related concepts. Tian et al. [21] showed a WS-QoS architecture that enables QoS-aware service specifications as well as the broker based Web service selection model that enables an efficient QoS-aware service selection. Eyhab Al-Masri and Qusay H. Mahmoud [22] introduced a mechanism that extends the Web Services Repository Builder (WSRB) of Web Services. It also introduced the Web Service Relevancy Function (WsRF) used for measuring the relevancy ranking of a particular Web service based on client s preferences and QoS metrics. Xu et al. [23] provided a Web service discovery model that contains an extended UDDI to accommodate the QoS information, a reputation management system to build and maintain service reputations and a discovery agent to facilitate service discovery. A service matching, ranking and selection algorithm is also developed. Demian Antony D Mello et al. [24] explored different types of requester s QoS requirements and a tree model for requester s QoS requirements. It also proposed a QoS broker based Web service architecture which facilitates the requester to select a suitable Web service based on QoS requirements and preferences. The Web service selection and ranking mechanism uses the QoS broker based architecture [26]. The QoS broker is responsible for the selection and ranking of functionally similar Web services. The Web service selection mechanism [26] ranks the Web services based on prospective levels of satisfaction of requester s QoS constraints and preferences. Serhani [27] proposed Web service architecture employs an extended UDDI registry to support service selection based on QoS, but only the certification approach is used to verify QoS and no information is provided about the QoS specification. Hongan Chen et al. [28] presented a description and an implementation of broker-based architecture for controlling QoS of Web services. The broker acts as an intermediary third party to make Web services selection and QoS negotiation on behalf of the client. Delegation of selection and negotiation raises trustworthiness issues mainly for clients. Performance of the broker is not considered in this approach. Moreover, performance of the broker can be a key to the success of any proposed architecture; if the user does not get a response to his/her request with an acceptable response time, he/she will switch to another provider. Wishart [29] offered a protocol for service reputation. An old factor for the reputation mark is used to each of the ranks for a service so current ranks are more important than aged ones. The value of the aged factor should rely on the requirements of the service consumers and the reputation management system. Hu [3] showed the Web service is selected by matching requested QoS property values against the potential Web service QoS property values. In this literature, the Web service is selected by taking the requester s average preference for QoS properties. Using FIPA compliant Multi Agents, Jaleh Shoshtarian Malak [34] proposed a Multi Agents based Web service QoS Management Architecture. A QoS based Web service clustering technique was introduced which helped in choosing a service that best suited user quality preferences. Many of these approaches do not provide guarantees as to the accuracy of the QoS values over time or having up-to-date QoS information. In the next section, we describe the design of the proposed Web Service broker-based architecture which overcomes many of the limitations imposed by current discovery model. Sri Sai Ram Engineering College, An ISO 91:28 Certified & NBA Accredited Engineering Institute, Chennai, INDIA. Page 3

III. International Journal of Emerging Technology and Advanced Engineering WSB WEB SERVICE ARCHITECTURE WITH QOS The purpose of Web service discovery is to select most suitable Web service for a particular task. When dynamic discovery is used in Web services, it is common that the result of the discovery contains more than one provider. We propose a technique for dynamic discovery of Web services which will also handle the problem of redundant Web services. The architecture consists of the basic Web service model components like the Web service provider, Web service consumer and the UDDI registry. In addition, UDDI registry has the capability to store QoS information using tmodel data structure and a WSB. Components of the architecture are presented in Fig. 1. We propose a Web services discovery architecture which contain an extended UDDI to accommodate the QoS parameters [23]. The WSB assists clients in selecting Web services based on a set of QoS parameters. The WSB has five components: Service Publisher [24], Verifier and Certifier, Retrieval Agent, Quality Analyzer, and Web Service Storage (WSS) [25]. Broker services may be used to facilitate service registry access. The broker performs the interaction with the UDDI. Fig. 1. Architecture for WSB The broker is a Web service performing a collection of QoS functionalities. It is the entity that performs the verification and certification tasks. It is also involved in other operations, such as registering and selecting services with QoS functions. To overcome many of the limitations imposed by current discovery model, we introduce the Web Service Broker architecture shown in Fig. 1. The service publisher component facilitates the registration, updating and deletion of Web service related information. It gets the business- and performancespecific QoS property values of Web services from the service providers. The service provider publishes its service functionality to the UDDI registry through the service publisher after certification and verification. For every service or group of services there exists a service publisher that handles all communication with registries, bindings, negotiations, requests, and responses for that service. The service consumer can search the UDDI registry for a specific service through the retrieval agent. The main functionality of the retrieval agent component is to select the most suitable Web service satisfying requester s QoS constraints and preferences, along with service functionality. The service requester can verify the advertised QoS with the retrieval agent before binding to a Web service. QoS verification is the process of validating the correctness of information described in the service interface as well as the described QoS parameters. The result of verification will be used as input for the certification process that will be issued when the verification succeed. The QoS property values obtained from the service providers are verified and certified by the Verifier and Certifier component before registering them into the UDDI registry. The Verifier and Certifier component is implemented within the WSB and is responsible for certifying Web services and their provided QoS. A certificate is sent to the Web service provider and a copy is stored in the WSS for future use. A sequence of interaction between these components is presented in Fig. 2. A typical usage scenario (Fig. 2) is described here by considering an example in which a consumer uses a Web service in its application. 1. Initially Web Service Broker (WSB) publishes the interface to the UDDI registry. 2. Web service provider finds the broker interface in UDDI registry. 3. The service provider registers the Web service with the service publisher which is available in the WSB and provides functional and non-functional information about the offered services. 4. The Verifier and Certifier component of the WSB verifies the QoS information and issues a certificate. 5. A copy of the QoS certificate is stored in WSS and a copy is sent to the service provider. 6. The service publisher then publishes the Web service in the UDDI registry along with the QoS certificate. 7. The consumer application requests service discovery and provides functional and QoS requirements. Sri Sai Ram Engineering College, An ISO 91:28 Certified & NBA Accredited Engineering Institute, Chennai, INDIA. Page 4

8. The retrieval agent finds the service in the UDDI registry according to the required service functionality and QoS requirements of the application. 9. The retrieval agent can communicate with quality analyzer to verify the provided QoS certificate and rating scores with the one stored in the WSS. 1. The retrieval agent then reports the discovered service back to the application (consumer). 11. The Web service consumer then binds the Web service from the service provider. 12. Consumer sends the feedback to the quality analyzer after invoked the service. 13. Quality analyzer calculates the rating scores for consumer feedback then stores it into database (WSS) for service discovery process. 1.1 Service Publisher The service publisher component communicates with the service provider and the UDDI registry. The service provider registers the business and Web service related information with the service publisher. It also gets the specific QoS property values of Web services from providers. Once the QoS property values and other information are obtained from the provider it is presented to the Verifier and Certifier component. The QoS information is verified and certified before publishing it in the UDDI registry. Fig. 2. Architectural Component Interactions 1.2 Verifier and Certifier This is the key component of the WSB that performs the verification of the QoS information supplied by the service provider and issues a certificate to the service provider through the service publisher. This QoS certificate assures that the QoS offered by the provider confirm to their descriptions. The service provider initiates the verification process through the service publisher by supplying the QoS property values. The verifier is provided with the WSDL document and additional information about resources available at the provider s platform. The verifier performs the testing of the service URI, the XML schema definition, the service binding information and the availability of all operations described in the service interface. Verifier also performs the verification of the QoS information introduced in the service interface. The QoS verification is conducted through a set of test cases generated by the verifier. For each test, additional information like server capacity, network bandwidth about the provider and its Web service are needed. The four QoS parameters (Response Time, Availability, Throughput, and Price) are also verified. The verification process is done in three levels: General Web services information verification, WSDL content verification, and QoS verification. A Web service is said to be compliant with a given level when it passes the corresponding tests described in the verification document. Based on this, Web services can be classified into three classes. Class A includes Web services for which all verification tests have succeeded. Class B includes Web services for which more than 8% of the verification tests have succeeded. Class C contains the services for which most of the verification scenarios have failed. Once the verification process is completed successfully, the certification process is initiated. The certifier issues a certificate to the service provider through the service publisher which indicates that the offered QoS confirm to their descriptions. The main responsibility of the certifier is to certify the Web services and their provided QoS. A copy of the certificate sent to the service provider, which is also stored in the WSS for future use. The certificate includes information such as certificate number, certificate issue date, number of years in business, class type, and service location. In case, if the certificate cannot be issued, feedback will be sent to the provider. After the QoS certification process, the service publisher can register the functional description of the service and the certified QoS information with the UDDI registry. Sri Sai Ram Engineering College, An ISO 91:28 Certified & NBA Accredited Engineering Institute, Chennai, INDIA. Page 5

1.3 Retrieval Agent The retrieval agent component is concerned with selecting the most suitable Web service satisfying the consumer s QoS constraints and the specific service functionality. It receives messages from the Web service consumer, specifying the service functionality along with the QoS constraints. Based on the received requirements specification, it discovers functionally similar Web services from the UDDI registry. The retrieval agent can check the validity of the QoS information in the UDDI registry by comparing the QoS certificate provided by the Verifier and Certifier with the one stored in the WSS. 1.4 Quality Analyzer After a Web service task is finished, a client may summarize the QoS experience and send them to the quality analyzer in WSB. The quality analyzer collects feedback regarding the service and QoS of the Web services from the service consumers, calculates rating scores, and updates these scores in the WSS. The quality analyzer uses this information during the service discovery phase. For this work, we assume that all ratings are available, objective and valid. Service consumers provide a rating, indicating the level of satisfaction with a service after each interaction with the service. A rating is simply an integer ranging from 1 to 1, where 1 shows extreme satisfaction and 1 shows extreme dissatisfaction. Every rating score includes service ID, consumer ID, a timestamp, and rating value. The service key in the UDDI registry of the service is used as the service ID, and the IP address of the consumer is used as the consumer ID. The quality analyzer stores rating score for services in a table. An example table is given in Table I. For example a consumer with IP address "192.168.2.5" for a service with key "673454-e8x6-11d5-9ea8-9325426754" provide rating score "9" in timestamp "211-3-1 1:15:9". Every consumer stores one rating for a service and future rating score for that service replaces previous rating. Only the most recent rating by a customer for a service is stored in the database. New ratings from the same customers for the same service replace older ratings. The time stamp is utilized to specify the previous factor of a special service rating. We use equation (1) in order to calculate the feedback rating. RS n n S i i 1 (1) RS means rating score for the service. Where n is number of rating for the service. S i means ith service rating for the service. Table I EXAMPLE RATINGS FOR SERVICES Service ID Consumer ID Timestamp Rating Score 673454-e8x6-11d5-192.168.2.5 211-3-1 9 9ea8-9325426754 1:15:9 9521db61-eac1-42e4-192.168.4.75 211-3-1 8 5ab-1d87b8f1876a 11:15:9 741549-fb-11d5- bca4-235229c64 192.168.2.7 211-3-2 11:25:21 7 921cb6e-e8c9-4fe3-9ea8-3c99b1fa8bf3 192.168.2.155 211-3-2 12:15:35 1.5 QoS Matching, Ranking and Selection Algorithm A Web service consumer sends a service discovery request to the retrieval agent, which then contacts the UDDI registry to find services that meet the customer s functional and QoS requirements. A service is said to be a match if it satisfies the customer s functional requirements and its QoS information. If no matched service is found by the matching process, the retrieval agent returns an empty result to the customer. If multiple services match the functional and QoS requirements, the retrieval agent calculates a QoS score for each matched service based on the dominant QoS attribute specified by the customer, or on the default dominant attribute, average response time. The best service is assigned a score of 1, and the other services are assigned scores based on the value of the dominant QoS attribute. The top M services (M being the maximum number of services to be returned as specified by the customer) with the highest QoS scores are returned to the customer. If M is not specified, one service is randomly selected from those services whose QoS score is greater than LowLimit. /*Web services matching, ranking and selection algorithm */ 1 findservices (functionrequirements, qosrequirements, feedbackrequirements, maxnumservices) { // find services that meet the functional requirements 2 fmatches = fmatch (functionrequirements); 3 if QoS requirements specified { // match services with QoS information 4 qmatches = qosmatch (fmatches, qosrequirements); } 5 else { // select max number of services to be returned 6 return selectservices (fmatches, maxnumservices, "random"); } 7 if feedback requirements specified { // matches with QoS and feedback information 8 matches = feedbackrank (qmatches, qosrequirements, feedbackrequirements); // select max number of services to be returned 9 return selectservices (matches, maxnumservices, "byqos"); } 1 else { // matches with QoS information 11 return selectservices (matches, maxnumservices, "byoverall"); } } Fig. 3. Service Matching, Ranking and Selection Algorithm Sri Sai Ram Engineering College, An ISO 91:28 Certified & NBA Accredited Engineering Institute, Chennai, INDIA. Page 6 6

Fig. 3 shows a simplified version of our service selection algorithm where the leftmost numbers denote the line numbers. When the retrieval agent receives a discovery request, it executes fmatch (line 2) which returns a list of services LS1 that meet the functional requirements. If QoS requirements are specified, qosmatch (line 4) is executed next on the set of services LS1 and it returns a subset of services LS2 that meet the QoS requirements. selectservices (line 6, 9, 11) always returns a list of M services to the customer where M denotes the maximum number of services to be returned as specified in the discovery request. If QoS requirements are not specified, selectservices returns M randomly selected services from LS1. If only one service satisfies the selection criteria, it returns this service to the customer. IV. EXPERIMENTAL RESULTS AND DISCUSSION This section presents experimental evaluation of the proposed approach. The proposed Web service broker based discovery system is implemented on Windows platform using Microsoft Visual Studio.NET development environment, Microsoft IIS 6. as a Web Server and ASP.NET, C#.NET as a programming language. To enable the interaction between.net program and the UDDI-compliant server, Microsoft UDDI SDK 2. is used. We used the database as Microsoft SQL Server, SQLExpress Edition 25. The QoS parameters considered for this approach are response time, service availability, price and throughput. In order to show the applicability of this approach, a prototype is developed. The experimental set up is carried out for four scenarios with the QoS properties such as response time, service availability, price and throughput. The QoS attributes are evaluated based on the number of clients and the available resources. 1.6 Scenario 1 (Discovery Process) Table 2 summarizes the QoS requirements of the 4 service consumers such as RT in milliseconds, TP in request per minute, price in dollar per transaction and service availability (AV) in percentage. All consumers specify that the maximum number of services to be returned is 1. All the four consumers C1, C2 C3 and C4 concentrates on the four QoS attributes namely response time, service availability, price and throughput. In this experimental set up, ten functionally matched online ticket booking services are considered and are returned from the UDDI. These services are selected by the WSB based on the QoS parameters such as response time, service availability, price and throughput. From these set of services, the WSB selects one best service and is returned to each consumer based on their requirements which are specified in Table 2. Consumer Table 2 CONSUMERS QOS REQUIREMENTS Requirements RT (ms) TP (req/min) Price(dollar /transaction) Sri Sai Ram Engineering College, An ISO 91:28 Certified & NBA Accredited Engineering Institute, Chennai, INDIA. Page 7 AV (%) Feedback Rating C1 965 9.4 97 7 C2 8 6.3 95 - C3 12 7.4 96 5 C4 11 5.3 98 - The values shown in Table 3 represent QoS values measured through the implemented prototype for ten different Web services. In order to find the most suitable Web service, it is essential to optimize the values for each QoS parameters. Table 3 QOS VALUES FOR VARIOUS AVAILABLE ONLINE TICKET BOOKING SERVICES Servic e ID Service Provide r Name RT(ms ) TP (req/min ) AV (% ) Price (dollars /transaction ) AC (% ) IN (% ) Feedbac k rating 1 SP1 85 8 97.4 92 95 7 1 1 8 2 SP2 71 12.5 96 3 SP3 9 7 95.1 95 94 6 1 1 9 4 SP4 7 11.3 98 5 SP5 1314 5 98.1 97 96 7 1 8 6 SP6 65 12 97.5 96 7 SP7 87 9 95.4 98 97 9 1 6 8 SP8 135 4.12 93 95 9 SP9 12 5 94.1 92 96 5 SP1 6 1 115 6 96.2 96 94 For each consumer, the same service discovery request was run 1 times and the service selected for each run is shown in Fig 4. For C1, C2, C3 and C4, a service is selected based on the response time, service availability, price and throughput which meet the consumer requirements. It is observed from the figure that the services 2, 4 and 6 are chosen randomly for C1 as the best services based on the requirements given in the Table 2. For C2, the services randomly selected by WSB are 2 and 4. Services 1, 2, 4, 6 and 7 are chosen randomly for the consumer C3 as it satisfies the requirements provided in Table 2. Finally, for the Consumer C4, the services randomly chosen are 2 and 4 based on the requirements.

Service selected 1 International Journal of Emerging Technology and Advanced Engineering 9 8 7 6 5 4 3 2 1 C1 C2 C3 C4 1 2 3 4 5 6 7 8 9 1 Discovery Request Sequence Fig. 4. Service Selection Validation of the other QoS attributes such as price and reputation is also considered for the service selection process. 1.7 Scenario 2 (Discovery Process) Table 4 shows the QoS values measured for seven Web services (Al-Masri and Mahmoud 27). The QoS parameter cost is represented in dollars and was provided by the service provider. In addition, response time, service availability, price and throughput values were measured over a period of time. In order to find the most suitable Web service, it is important to optimize the values for each QoS parameter. For instance, the Web service with lower response time is preferable than the Web service with higher response time. Table 4 QOS METRICS FOR VARIOUS AVAILABLE EMAIL VERIFICATION WEB SERVICES ID Service RT TP Price AV (%) Provider Name (ms) (req/min) (dollars/transaction) S XML Logic 1 Validate Email 72 6. 85.12 XWebservices S XWebEmail- 2 Validation 11 1.74 81.1 StrikeIron S Email 3 Verification 71 12 98.1 StrikeIron S Email Address 4 Validator 912 1 96.7 S CDYNE 5 Email Verifier 91 11 9.2 S Webservicex 6 ValidateEmail 1232 4 87 ServiceObjec S ts DOTS Email 7 Validation 391 9 99.5 Table 5 summarizes the QoS attributes such as RT in milliseconds, AV in percentage, TP in request per minute and price in dollar per transaction of 3 service consumers. All consumers specify that the maximum number of services to be returned is 1. C1 mainly focuses on the throughput and availability. When C2 and C3 are considered, it concentrates on the four QoS attributes. Consumer Table 5 CONSUMERS QOS REQUIREMENTS Requirements RT (ms) TP (req/min) AV (%) Price (dollars) C1-1 89 - C2 75 5 9.6 C3 11 7 99.1 Sri Sai Ram Engineering College, An ISO 91:28 Certified & NBA Accredited Engineering Institute, Chennai, INDIA. Page 8

Response Time (ms) International Journal of Emerging Technology and Advanced Engineering For each consumer, the same service discovery request was run 15 times and the service selected for each run is shown in Fig 5. For C1, a service is selected based on the throughput and availability. It is observed from the figure that the services 3, 4 and 5 are chosen randomly as the best services based on the requirements given in the Table 5. For C2, the randomly selected services are 3 and 5. Service 3 is chosen for the consumer C3 as it satisfies the requirements provided in Table 5. Table 6 EVALUATION OF RESPONSE TIME FOR ONLINE TICKET BOOKING SERVICE Number of Clients Response Time (ms) 2 4 4 512 1 75 14 15 18 1498 22 1654 26 2145 4 3 Average Response Time 2 Fig. 5. Service Selection 1.8 Scenario 3 (Verification Process for Registration) This section presents the scenario 3 of the experimental results. This simulation model comprises of a single broker, a Web service and N concurrent clients. This section mainly focuses on the service registration. The QoS parameters taken into consideration are response time, throughput and service availability. The following test cases are carried out to verify the QoS properties. Consider a service provider which would provide service for the registration process. The proposed WSB approach evaluates the given the services and issues certificate or send appropriate feedback to the provider based on the verification results. 1.8.1 Response Time Table 6 shows the response time observation by the proposed WSB architecture. The values are tabulated for varying the number of clients. From the Table, it is observed that the response time varies with the different number of clients. With increase in the number of clients, the response time also increases. 1 2 4 1 14 18 22 26 Number of Clients Fig. 6. Distribution of RT with Increased Number of Clients Fig 6 illustrates the graphical representation of the Response Time for the online ticket booking service. It is observed from the Fig 6 that the response time increases linearly with the number of clients. For instance, when the number of clients is 26, the response time taken is 2145 milliseconds. 1.8.2 Service Availability Table 7 shows the service availability evaluation of the proposed WSB approach. Initially, the service availability is steady and it is 1% till 3 clients. Then the service availability fluctuates. Then, when the number of client become 1, the service availability becomes steady and it reaches 1% until it reaches the fluctuation points where the service availability decreases for 22 and 24 clients. Sri Sai Ram Engineering College, An ISO 91:28 Certified & NBA Accredited Engineering Institute, Chennai, INDIA. Page 9

Service Availability (%) International Journal of Emerging Technology and Advanced Engineering Table 7 EVALUATION OF SERVICE AVAILABILITY FOR ONLINE TICKET BOOKING SERVICE Table 8 EVALUATION OF THROUGHPUT FOR ONLINE TICKET BOOKING SERVICE Number of Clients Service Availability (%) 2 1 3 1 4 97 5 98 1 1 14 1 18 1 22 94 26 9 The graphical representation of the service availability of the proposed broker based approach is shown in Fig 7. 1 Number of Clients Throughput (req/min) 2 17 3 17 4 15 5 13 1 11 14 8 18 6 22 5 26 4 Fig 8 illustrates the graphical representation of the throughput request per minute for online ticket booking service. 95 9 85 Service Availability 8 2 3 4 5 1 14 18 22 26 Fig. 7. Distribution of Service Availability with Increased Number of Clients 1.8.3 Throughput Table 8 shows the throughput values of online ticket booking service with increasing number of clients. When the number of clients increases, there is slight decrease in the throughput values. Fig. 8. Distribution of Throughput with Increased Number of Clients The results of the scenario 3 show that the values of Response time, service availability and throughput are fluctuating with various numbers of clients. Thus, it is concluded from this scenario that the number of client have a significant effect on response time, availability of the service and throughput. Sri Sai Ram Engineering College, An ISO 91:28 Certified & NBA Accredited Engineering Institute, Chennai, INDIA. Page 1

Finally, the results of the validation test cases show significant impact on the server resources capacity, the number of connected client, the network load on response time, throughput and the availability of the Web service. 1.9 Scenario 4 (Verification Process for Registration) The functionally matched service obtained from the UDDI registry is verified by the Web Service Broker based on the QoS parameters such as response time, service availability, price and throughput. 1.9.1 Response Time Response time of the online ticket booking service is evaluated in Fig 9. The Figure illustrates the graphical representation of the response time for 1 discovery request sequence. It is observed from the Fig 9 that the response time lies within the range of 7 to 78 milliseconds for all the ten discovery request sequence. The highest response time taken by the given ticket booking service is 773 milliseconds. The average response time taken by the given online ticket booking service is 758.8 milliseconds. The average response time of the ticket booking service is compared with the service response time which is given in the tmodel data structure in the UDDI Registry. If the response time is matched with the values in the tmodel then, the service would be selected. If the value is not matched up with the values in the UDDI registry, then the service would not be selected by the WSB. 1.9.2 Throughput The throughput results of the online ticket booking service are shown in Fig. 1. The throughput is obtained for the ten discovery request sequences. The highest throughput obtained for ticket booking service is 12 requests per minute. The obtained throughput values of the ticket booking service from WSB are compared with the throughput values which are given in the tmodel data structure in the UDDI Registry. If the throughput is matched with the values in the tmodel then, the service would be selected. If the value is not matched up with the values in the UDDI registry, then the service would not be selected by the WSB. Fig. 1. Throughput Verification Fig. 9. Response Time Verification 1.9.3 Service Availability Similarly, Figure 4.13 shows the service availability verification process. The service availability is obtained for the ten discovery request sequences. The highest service availability obtained for ticket booking service is 1%. For most of the discovery request sequences, the service availability is 1%. For example, for the discovery request sequence such as 1, 2, 3, 5, 6, 8 and 1, the service availability obtained is 1%. The obtained service availability values of the ticket booking service from WSB are compared with the service availability values which are given in the tmodel data structure in the UDDI Registry. If the service availability is matched with the values in the tmodel then, the service would be selected. If the value is not matched up with the values in the UDDI registry, then the service would not be selected by the WSB. Sri Sai Ram Engineering College, An ISO 91:28 Certified & NBA Accredited Engineering Institute, Chennai, INDIA. Page 11

The key features of the broker that are not supported by existing approaches dealing with QoS for Web services are focused in this approach. The service provider does not have to design and develop her/his own broker but just invoke one from the already published brokers from the Internet. The client will also find a good support during its Web services discovery using the broker services. This will enable a more flexible, and trustable architecture. The experimental results are carried out in two scenarios with QoS attributes such as Response Time, Service Availability and Throughput. The experimentation results confirmed the importance of QoS in distinguishing the functionally similar Web services. Fig. 11. Service Availability Verification After the verification process, the services which are all matched with QoS requirements given in the interface are selected by the WSB. Then, the best services selected by the WSB are given to the consumer. The proposed WSB will exploit the results of the above experiments to evaluate the response time, service availability and throughput of the service to its provider. Validation of the QoS attributes such as response time, service availability and throughput is achieved by the WSB for online ticket booking service. Then, the verifier analyses the results obtained. Verifier also performs the verification of the QoS information given in the service interface. The service verifier store all important QoS information published in the service interface in its database. Once the verification process is completed successfully, the certification process is initiated. The certifier issues a certificate to the service provider through the service publisher which indicates that the offered QoS confirm to their descriptions. If the evaluated results of online ticket booking service do not meet the specified QoS descriptions in the service interface, then the certifier provides an appropriate feedback to the service provider through the service publisher. 1.1 Summary This section explores a broker-based architecture for Web services registration and discovery with QoS. The goal of the broker is to support Web services discovery with QoS registration, verification, certification and confirmation. The broker performs the process of publishing and selection of Web services. V. CONCLUSION In this paper, the WSB performs the process of registering and selection of Web services. The key features of the broker that are not supported by existing approaches dealing with QoS for web services are described. The service provider does not have to design and develop her/his own broker but just invoke one from the published brokers. The client will also find a good support during its Web services discovery using the broker services. The goal of the broker is to support Web services discovery with QoS registration, verification, certification, and confirmation. This approach helps in identifying the most suitable Web service according to the consumer s requirements. Security techniques can be included in the proposed approaches for a more flexible and trustable architecture. Another future work would be to enhance the adaptability of the present Web service architecture to support mobile Web services and also to enhance the capabilities of the proposed architecture to handle other QoS attributes. REFERENCES [1] S. Ran, A Model for Web Services Discovery with QoS, ACM SIGEcom Exchanges, Vol. 4(1), pp.1 1, 23. [2] W3C, QoS for Web Services: Requirements and Possible Approaches. Available: http://www.w3c.or.kr/kroffice/tr/23/note-ws-qos-231125/, 23. [3] IBM Corporation, Web Service Level Agreement (WSLA) Language Specification Ver. 1.. Retrieved from http://www.research.ibm.com/wsla/wslaspecv1-23128.pdf, 23. [4] W3C, WS-Policy Framework ver.1.2, Available at: http://www.w3.org/submission/ws-policy/, 26. [5] DAML Services, DAML-S / OWL-S, Available at: http://www.daml.org/services/owl-s/, 26. [6] E.M. Maximilien and M.P. Singh, Reputation and Endorsement for Web Services, ACM SIGecom Exchanges, Vol. 3(1), pp.24 31, 22. [7] E.M. Maximilien and M.P. Singh, Self-Adjusting Trust and Selection for Web Services, In extended Proc. Of 2nd IEEE Intl. conf. on Autonomic Computing (ICAC), pp.385-386, 25. Sri Sai Ram Engineering College, An ISO 91:28 Certified & NBA Accredited Engineering Institute, Chennai, INDIA. Page 12

[8] L. Vu, M. Hauswirth and K. Aberer, QoS-based service selection and ranking with trust and reputation management. In Proc. of the Intl. Conf. on Cooperative Information Systems (CoopIS), Agia Napa, Cyprus, 25. [9] E.M. Maximilien and M.P. Singh, Towards Autonomic Web Services, Trust and Selection, ICSOC 4, pp.212 221, November 24. [1] A. Blum, UDDI as an Extended Web Services Registry: Versioning, quality of service, and more, White paper, SOA World magazine, Vol. 4(6), 24. [11] Shrabani Mallick and D.S. Kushwaha, An Efficient Web service Discovery Architecture, International Journal of Computer Applications, Volume 3-No 12, July 21. [12] S. Majitha, A. Shaikhali, O. Rana, and D. Walker, Reputation based semantic service Discovery, In Proc. Of the 13 th IEEE Intl Workshops on Enabling Technologies Infrastructures for collaborative Enterprises (WETICE), Modena, Italy, pp.297-32, 24. [13] T. Rajendran and P. Balasubramanie, An Efficient Framework for Agent-Based Quality Driven Web Services Discovery, IEEE International conference on Intelligent Agent and Multi Agent Systems (IAMA29), Chennai, 29. [14] A. Keller and H. Ludwing. The WSLA framework: Specifying and Monitoring Service Level Agreements for Web Services, IBM Research Report, 22. [15] P. Farkas and H. Charaf, Web Services Planning Concepts, 1 st International Workshop on C# and.net Technologies on Algorithms, Computer Graphics, Visualization, Distributed and WEB Computing, Feb. 23 [16] A. ShaikAli, O.F. Rana, R. Al-Ali, and D.W. Walker, UDDIe: An extended registry for Web services, In Proc. Of the Symposium on Applications and the Internet workshops, IEEE CS, pp.85-89, 23. [17] Z. Chen, C. Liang-Tien, B. Silverajan, and L. Bu-Sung, UX-An architecture providing QoS-aware and federated support for UDDI, In Proc. of the Int l Conf. on Web services, CSREA Press, pp.171-176, 23. [18] S. Kalepu, S. Krishnaswamy, and S.W. Loke, Reputation = f (User Ranking, Compliance, Verity), Proceedings of ICWS'4, 24. [19] Y. Liu, A. Ngu, and L. Zheng, QoS Computation and Policing in Dynamic Web Service Selection, Proceedings of WWW 24 Conf, 24. [2] Diego Zuquim Guimaraes Garcia and Maria Beatriz Felgar de Toledo, A Web service Architecture providing QoS Management, Web Congress, LA-Web '6, Fourth Latin American, pp -189-198, 26. [21] M. Tian, A. Gramm, H. Ritter, and J. Schiller, Efficient Selection and Monitoring of QoS aware Web Services with the WS-QoS Framework, Proceedings of the IEEE/WIC/ACM International Conference on Web Intelligence (WI 4) Exchanges, vol. 4, no. 1, pp. 1 1, 24. [22] Eyhab Al- Masri and Qusay H. Mahmoud, QoS-based Discovery and Ranking of Web services, Proceedings of IEEE International Conference, 27. [23] Ziqiang Xu, Patrick Martin, Wendy Powley, and Farhana Zulkernine, Reputation Enhanced QoS-based Web services Discovery, IEEE International Conference on Web Services (ICWS 27), 27. [24] Demian Antony D Mello, V.S. Ananthanarayana, and T. Santhi, A QoS Broker Based Architecture for Web Service Selection, Proceedings of IEEE International Conference, 28. [25] T. Rajendran and P. Balasubramanie, An Agent-Based Dynamic Web Service Discovery Framework with QoS Support, International J. of Engg. Research & Indu. Appls. (IJERIA), Vol. 2(5), pp 1-13, 29. [26] Demian A. D Mello and V.S. Ananthanarayana, A QoS Model and Selection Mechanism for QoS-Aware Web Services, Proceedings of the International Conference on Data Management (ICDM 28), 28. [27] M.A. Serhani, R. Dssouli, A. Hafid, and H. Sahraoui, A QoS broker based architecture for efficient Web services selection. In Proc. of the IEEE Int l Conf. on Web Services, IEEE CS, pp.113 12, 25. [28] Hongan Chen, Tao Yu, and Kwei-Jay Lin, QCWS: an implementation of QoS-capable multimedia Web services, IEEE Fifth International Symposium on Multimedia Software Engineering, 23. [29] R. Wishart, R. Robinson, J. Indulska, and A. Josang, SuperstringRep: Reputation-enhanced Service Discovery. In Proc. of the 28th Australasian conf. on Computer Science, Vol. 38, pp.49-57, 25. [3] J. Hu, C. Guo, H. Wang, and P. Zou, 25, Quality Driven Web Services Selection Proceedings of the IEEE International Conference on e-business Engineering (ICEBE 5), 25. [31] M. Crasso, J.M. Rodriguez, A. Zunino, and M. Campo, Improving Web service descriptions for effective service discovery, Sci. Comp. Prog., 75, pp.11-121, 21. [32] Vuong Xuan Tran, Hidekazu Tsu ji, and Ryosuke Masuda, A new QoS ontology and its QoS-based ranking algorithm for Web services, In Elsevier Journal: Simulation Modelling Practice and Thoery, 29. [33] Mydhili K. Nair and Dr. V. Gopalakrishna, Look Before You Leap: A Survey of Web Service Discovery, International Journal of Computer Applications, Volume 7, No.5, pp.22-3, September 21. [34] Jaleh Shoshtarian Malak, Mohsenzadeh, M. and Seyyedi, M.A. Multi agent based Web Service QoS management architecture, 14th International CSI Computer Conference (CSICC), pp. 28 286,29. Sri Sai Ram Engineering College, An ISO 91:28 Certified & NBA Accredited Engineering Institute, Chennai, INDIA. Page 13