Biography

Design of a Communication Infrastructure for GUISET Services based on Multiple Enterprise Service Buses

Categories
Published
of 120
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Description
Design of a Communication Infrastructure for GUISET Services based on Multiple Enterprise Service Buses Themba Shezi ( ) A dissertation submitted in fulfilment of the requirements for the degree
Transcript
Design of a Communication Infrastructure for GUISET Services based on Multiple Enterprise Service Buses Themba Shezi ( ) A dissertation submitted in fulfilment of the requirements for the degree of Master of Science in Computer Science Department of Computer Science, Faculty of Science and Agriculture University of Zululand KwaDlangezwa 3886 RSA Supervisor: E. Jembere Co-Supervisor: Dr. J Oladosu 2013 DECLARATION I, Themba Shezi, declare that this dissertation represents my own work and that this work has not been previously submitted at any university or other institution of tertiary education. All sources of information used in this work have been acknowledged. Themba Shezi Signature i DEDICATION To my late father, John Shezi ii ACKNOWLEDGEMENT First, I would like to thank the Almighty God, without his enduring mercy I could not have achieved anything. Secondly, I would like to express my gratitude to supervisors Mr E. Jembere and Dr J Oladosu for their support and guidance throughout to make this research work a reality. I was very fortunate to have them as my supervisors. I would also like to extend my special thanks to my Prof. M.O Adigun and Prof. S.S Xulu for their fatherly support. I would also like to thank my fellow researchers and friends L. Nkosi, S.C Makhaye, S.K.S Ngwenya, M. T Nene,S. W Dlamini, S. Cebekhulu, P.T. Cwele and N.M Gumbi, thank you for all the support and motivation you gave me during the course of this work. Thanks to all members of the Centre for their support and assistance. In particular, P. Mudali, P.Tarwirei, B. Mutanga, A. Alaba, E. K Olatunji, M.V Shabalala, S. Fatyi, O Kayode, N. Sibeko, Z. Ndlela and N. Mdletshe. I would also like to thank my mother and the whole family for their love, understanding and support. Last, but not least, I would like to express my sincere appreciation to Miss P. Z Gumbi for her support throughout this study. iii TABLE OF CONTENTS DECLARATION... i DEDICATION... ii ACKNOWLEDGEMENT... iii ABSTRACT... xii CHAPTER ONE... 1 INTRODUCTION Introductory Background Statement of the Problem Research Goal Research Objectives Research methodology Research Design Research Methods and Techniques Dissertation Synopsis CHAPTER TWO BACKGROUND OF ENTERPRISE SERVICE BUS CONCEPT Introduction The GUISET Project The GUISET Architecture GUISET Integration Requirements Enterprise Integration Enterprise Resource Planning (ERP) Enterprise Application Integration (EAI) Enterprise Service Bus (ESB) Service Oriented Architecture (SOA) The Web Services SOAP Web Services Description Language (WSDL) Service Discovery UDDI Service registry iv 2.4.6 Service Selection Web Service Composition Enterprise Service Bus and Service Oriented Architecture Comparison of Enterprise Integration Approaches ESB Solutions Assessing the ESBs Overview of the selected ESBs JBoss ESB Mule ESB Apache ServiceMix ESB Chapter Summary CHAPTER THREE THEORETICAL EVALUATION OF ESBs Introduction General ESB Evaluations Core GUISET ESB Evaluation ESB High Availability Data Transformation Intelligent Message Routing (Content-based routing) Dynamic Service Discovery Service Composition Analysis Methodology Pair-wise comparison process Priority Assignment and Judgments The Overall Criteria Rankings The Overall ESB Ranking A discussion on Analysis of Results Chapter Summary CHAPTER FOUR EMPIRICAL EVALUATIONOF ESBs Introduction v 4.2. ESB with Service Orchestration BPEL Engine and WS-BPEL ESB and UDDI The UDDI Registry Dynamic service discovery and Selection Motivating Scenario: Loan Broker Application Implementation Overview Walk-through of the Loan Broker Application Integration Model Design for our ESB Performance Evaluation Simple ESB Direct Service Orchestration ESB integrated with BPEL engine for service orchestration ESB integrated with UDDI for dynamic service discovery Performance Evaluation of ESBs Scalability Average Response Time Throughput Statistical Analysis Method Experimental Setup and Results Analysis Chapter Summary CHAPTER FIVE ESB FEDERATION PATTERNS ANDPERFORMANCE EVALUATION Introduction Directly Connected ESB Federation Pattern Brokered ESB Federation pattern Hub and Spoke ESB Federation Pattern Design of Federated ESB Implementation of LoanBroker Scenario Apache juddi Registry hosting LoanBroker Services BPEL LoanBroker Processes Implementation of Federated ESB Patterns Directly Connected ESB Federation Pattern vi Hub and Spokes ESB Federation Pattern Brokered ESB Federation Patterns Performance Evaluation of Federated ESB Patterns Basic Assumption of the Simulation Model Simulation Setup and Environment Experimental Results and Discussion Results Discussion Performance Comparison of a single ESB and Directly Connected ESBs Configuration of each ESB with Apache ODE and juddi Performance evaluation of ESB integrated with UDDI and BPEL Engine Comparing Performance Resultsfor ServiceMix and Directly Connected ESB Increasing number of services discovered Increasing number of services published Increasing number of concurrent requests with respect services published Results Discussion for ServiceMix and Directly Connected ESBs Chapter Summary CHAPTER SIX SUMMARY AND FUTURE DIRECTIONS Summary Limitations and Future Directions REFERENCES Appendix A: ESB Configuration and Installations Appendix B: Source Code BPEL LoanBroker Process Implemented for ESB Federation Dynamic Service Discovery Mechanism vii LIST OF FIGURES Figure 1.1: The path from business agility to the ESB (Cape Clear, 2005)... 3 Figure 2.1: GUISET Architecture (Adigun et al., 2006) Figure 2.2: SOA publish-find-bind architecture Figure 2.3: Web service standards (Papazoglou, 2008) Figure 2.4: UDDI core data structure (Clement et al., 2004) Figure 2.5: QoS information stored on UDDI tmodel data structure (Blum and Fred, 2004) Figure 2.6: BPEL Components Figure 2.7: Sample BPEL process Figure 2.8: JBoss ESB Architecture using example Figure 2.9: Mule ESB Architecture Figure 2.10: Apache ServiceMix ESB Architecture Figure 3.1: AHP steps for Analysis of ESBs for GUISET Figure 3.2: Average weights and Ranking for all criteria Figure 4.1: LoanBroker Sequence diagram Figure 4.2: ServiceMix configuration for Direct Service Orchestration Figure 4.3: Mule configuration for Direct Service Orchestration Figure 4.4: JBoss configuration for Direct Service Orchestration Figure 4.5: ServiceMix configuration for BPEL Orchestration Figure 4.6: Mule configuration for BPEL Orchestration Figure 4.7: JBoss configuration for BPEL Orchestration Figure 4.8: Overview of ESB integrated with UDDI Figure 4.9: ServiceMix ESB integrated with UDDI Figure 4.10: Mule ESB integrated with juddi Figure 4.11: JBoss ESB integrated with UDDI Figure 4.12: Response Time vs. No of Requests. 84 Figure 4.13: Throughput vs. No of Requests Figure 4.14: Response time vs. No of Requests 85 Figure 4.15: Throughput vs. No of Requests Figure 4.16: Response Time vs. No of services discovered..88 Figure 4.17: Throughput vs. No of services discovered Figure 4.18: Response time vs. No of services published.89 Figure 4.19: Throughput vs. No of services published Figure 4.20: Response time vs. no of requests..90 Figure 4.21: Throughput vs. no of requests viii Figure 5.1: Overview Design solution for Federated ESB Figure 5.2: Service Provider Component Figure 5.3: Interaction between Federated ESB and other Components Figure 5.4: BPEL Engine Component Figure 5.5: The UDDI Registry Component Figure 5.6: Snapshot of the Apache juddi showing LoanBroker services published Figure 5.7: BPEL defined LoanBroker process Figure 5.8: Directly Connected ESB Federation implements of LoanBroker Figure 5.9: Hub-Spoke Federation pattern implements LoanBroker Figure 5.10: Brokered ESB Federation Pattern implements LoanBroker Figure 5.11: Response Time vs. No of service discovered.112 Figure 5.12: Throughput vs. No of service discovered Figure 5.13: Response Time vs. No of service published Figure 5.14: Throughput vs. No of service published Figure 5.15: Response time vs. increasing no of requests Figure 5.16: Throughput vs. increasing no of requests Figure 5.17: Mule Integrated with UDDI and BPEL Engine Figure 5.18: ServiceMix ESB integrated with UDDI and BPEL Engine Figure 5.19: JBoss ESB integrated with UDDI and BPEL Engine Figure 5.20: Response time vs. No. of services discovered Figure 5.21: Throughput vs. No. of services discovered Figure 5.22: Response time vs. no of services published 112 Figure 5.23: Throughput vs. no of services published Figure 5.24: Response time vs. No of Requests 122 Figure 5.25: Throughput vs. No of Requests Figure 5.26: Response time vs. no of services discovered..127 Figure 5.27: Throughput vs. no of services discovered Figure 5.28: Response time vs. no of services published Figure 5.29: Throughput vs. no of services published Figure 5.30: Response time vs. increasing no of requests Figure 5.31: Throughput vs. increasing no of requests ix LIST OF TABLES Table 2.1: Comparisons of Enterprise Integration pattern Table 2.2: ESB evaluations against general criteria Table 3.1: Intensity Scale and Definition Table 3.2: ESB technologies towards supporting GUISET integration requirements Table 3.3: Pairwise comparison matrix for High Availability (HA) Table 3.4: Normalized table for High availability Table 3.5: Average Random Consistency (Al-Harbi, 2001) Table 3.6: Pairwise Comparison matrix for the remaining criteria Table 3.7: Pairwise comparison matrix for all five criteria Table 3.8: Overall ESB Rankings Table 4.1: P-Values for Direct and BPEL Service orchestration Table 4.2: Means and grouping of ESBs for Direct and BPEL Service Orchestration Table 4.3: P-Values for Dynamic Service Discovery Table 4.4: Means and grouping of ESBs for Dynamic Service discovery Table 5.1: P-Values for comparison of ESB Federation patterns Table 5.2: Means and grouping for ESB Federation patterns Table 5.3: P-Values for ESB integrated with UDDI and BPEL Engine Table 5.4: Means and grouping of ESBs integrated with UDDI and BPEL Engine Table 5.5: P-Values for Directly Connected ESBs and ServiceMix ESB Table 5.6: Computed mean values for Directly Connected ESBs and ServiceMix ESB x List ofpublications Shezi T., Jembere E., Adigun M.O., Nene M.T. (2013). Enabling Dynamic Service Discovery and Composition in the Enterprise Service Bus. In Proceedings of SATNAC 2013 Conference, Stellenbosch, September Shezi T., Jembere E., Adigun M.O., Nene M.T. (2012). Analysis Of Open Source Enterprise Service Buses Toward Supporting Integration In Dynamic Service Oriented Environments. In Springer Link (LNICST series)for EAI Conference on e Infrastructure and e Services for Developing Countries Shezi T., Jembere E., Adigun M.O. (2012). Performance Evaluation of Enterprise Service Buses Towards Support of Service Orchestration. Proceedings, PSR Centre: International Conference on Computer Engineering and Network Security (ICCENS), Dec 26-27, 2012, Dubai, pp Shezi T., Jembere E., Adigun M.O (2011). Towards Developing Failure Tolerant Communication Framework for GUISET Services. In Proceedings of SATNAC Conference, East London, Sept xi ABSTRACT In recent years, Service Oriented Architecture (SOA) has become a paradigm for enabling more efficient and flexible business processes in a service-based economy. The significance of this paradigm results in many organizations moving their businesses and making them available as online services so that they can be accessed ubiquitously by anyone connected to the network. The idea is to increase level of resource sharing and collaboration among geographically dispersed individuals/organizations. One of the successful SOA implementation that has recently received a lot of attention is the Enterprise Service Bus (ESB). ESB provides a key infrastructure that support guaranteed event handling, durable messaging, and data transformation capabilities that are needed by SOA environments. The success of ESB resulted in many ESB products being implemented and offered as both commercial and open source integration solutions. However, these products offer different approaches towards achieving ESB capabilities. Therefore, selecting the most suitable ESB becomes a challenging task, not only because there are many factors to consider in this selection, but also owing to the relationships between these factors and requirements of a particular integration scenario. There are many research efforts that have attempted to assist in ESB selection. They only consider evaluation of ESB products against given integration requirements. These evaluations are only useful when there is an ESB product that best support all the integration requirements of a given environment. This is hardly the case because ESBs perform well in some capabilities and worst in others. It is, therefore, believed that multiple ESBs can be integrated to get the best of individual ESBs that can give better performance compared to a single ESB. On the backdrop of the foregoing, this work considered GUISET integration requirements and investigated the validity of the above mentioned belief by integrating multiple ESBs to work together as a xii federation. Federation of ESBs allows each ESB to be used for the capability it best supports. Key capabilities investigated in this study are Service Discovery and Composition. An investigation was carried to find out which among the three (3) ESBs (ServiceMix, Mule and JBoss) considered best supports each of the afore-mentioned capabilities. The results showed that ServiceMix has the best support for Service Composition while JBoss has the best support for Service Discovery. These findings were then used for empirical evaluation of Directly Connected, Hub-Spoke and Brokered ESB Federation patterns, with each ESB providing the capability it best supports to the federation. Directly Connected ESB federation pattern outperformed the other patterns. We then compared the performance of Directly Connected ESB Federation and ServiceMix ESB to determine whether ESB federation has better performance compared to a single ESB. The results showed that ESB federation has better performance in terms of response time and throughput compared to a single ESB. xiii CHAPTER ONE INTRODUCTION 1.1 Introductory Background In today s world, businesses either small or large are striving for growth and competitiveness in order to be on the cutting edge. As a results business requirements and processes change frequently. To retain its agility and adaptability, businesses must ensure that their supporting IT systems and resources respond quickly to the changing business needs. The most basic needs of any business include cutting costs, quicker response to customer needs, integration of different systems across, the organization collaboration with other businesses through Business to Business (B2B) and Business to Consumer (B2C) interactions and achieving greater Return on Investment (ROI) (Fakorede,2007). In response to the dynamic business needs, companies are required to build technologies from scratch, integrate applications across incompatible platforms and manually coordinate business process execution. None of these solutions are efficient because they are prone to error, very costly, difficult to maintain and could result to duplication of effort. While this is the case, business climate requires business processes to be automated to enable easy composition and integration of different systems. In addition, systems must be developed in such a way that they are supported across heterogeneous environment such as various operating systems, hardware, programing languages and middleware. One solution that meets these requirements is Service Oriented Architecture (SOA) (Papazoglu and Heuvel, 2007). SOA has become a paradigm that allows interoperability between heterogeneous systems. SOA is acknowledged as architecture for integrating more complex systems at reduced cost and developing loosely coupled applications by means of services. Services are self-contained, well 1 defined and independent software component that defines a certain business function. Web service standards are the key enabling technologies for SOA. These standards include Web Service Definition Language (WSDL) which is an XML-based language for defining web services, Universal Description, Discovery and Integration (UDDI), which act as a repository for storing information about web services and SOAP which defines message structure for communication exchange over the Internet. The significance of SOA paradigm in enabling efficient and effective business processes in the service-based economy results in many organizations moving their businesses to making them available as online services so that they can be accessed ubiquitously by any entity connected to the network (Serhani, 2010). The idea is to increase level of resource sharing and collaboration among geographically dispersed individuals and organizations. However this shift has partially reached Small, Medium and Micro Enterprises (SMMEs) especially in African countries due to the cost associated with the underlying infrastructure that realize this new way of doing business (Adigun et al., 2006; Bauler et al., 2006; Zdravkovic et al., 2007). Therefore, to address this issue and others, an open dynamic service oriented and grid environment called Grid-based Utility Infrastructure for SMME Enabling Technologies (GUISET) was proposed in (Adigun et al., 2006). GUISET aims to leverage on the existing computing paradigm such as Service Oriented, Grid and Utility computing to provide affordable technologies to SMMEs. The idea is to provide a platform that would allow SMMEs to share information and resources thereby helping them to sell their products online and increase their trading territory without spending much on the technology. Chapter Two presents more details about GUISET. 2 The SOA implementation based on endpoints alone falls short of the key enabling infrastructure that support data transformation, guaranteed event handling, durable messaging and orchestration of multiple services using workflows (Genender, 2006). The importance of these additional requirements led to the concept of Enterprise Service Bus (ESB).Figure 1.1 shows the path from business needs to the ESB product. As previously mentioned, SOA is an architecture that delivers flexible, interoperable and cost saving approach towards achieving business needs. While, on the other hand, Web service is the key enabling technology for SOA architecture. ESB is an open, standard-based integration infrastructure designed to enable the implementation, deployment and management of SOA-based solutions with a focus on assembling, deploying and managing distributed SOA (Papazoglu and Heuvel, 2007). Figure 1.1: The path from business agility to the ESB (Cape Clear, 2005) ESB integrate heterogeneous applications by providing a standard-based integration platform that combines web service, messaging routing, data transformation and service virtualization. Message routing is the core capability of ESB and it is based upon a number of factors, such as message content, message header and transport type. Data transformation is another capability of 3 ESB that enables data to be changed from source format to the one required by the destination application. Data transformation support loose coupling of communicating applications. Thus an ESB is responsible for transporting data, transforming it and routing it to the appropriate endpoint service (Goldshlager and Zhang, 2005; Ziyaeva et al., 2008).The industrial success of ESB technology resulted in many products being implemented and offered as both commercial and open source products. Commercial products include Biztalk server, BEA Aquatic Service Bus, IBM WebSphere, Tibco, and Oracle ESB. Open source market includes Mule, ServiceMx, WSO2, JBoss, Petals and Open ESB (Vollmer et al., 2011). These product imp
Search
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks