[中文]
建立在基因算法上的QoS感知网络服务选择如何解决网络服务选择的结构性难题
摘要:SOA(服务导向架构)自出现以来已经成为把网络应用从简单的应用元素服务到建立复杂式分布服务的典范。但是,随着全球范围内网络服务供应商数量的激增,越来越多的同等功能服务的数量也呈百花齐放之势。于是,面对类型繁杂的技术一种建立在客户QoS要求基础上QoS感知网络服务选择被提了出来。在这篇论文里,为了能够在众多的同等类型的服务中条学出在核实的服务,我们提倡QoS感知网络服务选择法。实验结果表明,我们所提倡的这种方法对于解决网络服务的构成问题是极为有效的。
关键词:QoS(服务质量),网络服务,服务选择,基因算法
I.论文简介
大家都知道,网络是一个巨大的分散式的动态的信息库。现在网络已演变到可以囊括世界范围内的各种的可用的信息资源的程度。于是网络服务就应运而生了。网络服务是一种可以自我解释的软件实体,可根据一系列的标准(如SOAP, WSDL, 和 UDDI 等)通过互联网提供广告,定位和使用等服务【1】。网络服务封装功能应用和信息资源,保证通过促使编程性通路在整个网络的应用,进一步加强网际应用程序交互作用的力度。
在同一个网络环境里,多样化的网络服务可以提供多种同等功能的服务,往往还带有不同的非功能性的属性值。这时,这些网络服务就会在一个单元里拥挤在一起,这是很典型的一种情况。为了在选择服务时区分同一单元里不同的服务类型,我们要考虑他们的QoS价值,我们在这里也强烈建议在选择网络服务时把它们的QoS价值考虑进去。
.大家都知道,对于系统整合商所面临的寻址平台的互动性和兼容性的问题,网络服务技术是一很有前景的解决方案。但是,由于在网络服务的选择过程缺乏QoS,它们的采用速率很低。网络服务技术仍然面临:“我怎样才能知道这项网络服务否满足我的绩效需要?”诸如此类的问题。若不解决以上问题,普遍的描述、发现和集成(UDDI)注册表的商业应用无异于天方夜谭此外。另一个极其重要且富有挑战性的问题是,网络上不同QoS标准的同等功能的服务正在增加。根据【2】数据显示,2000~2000年10月期间,已发行网络服务增幅超过130%。由网络服务搜索引擎Seekda所发布的数据同样也显示在过去的三年里,网络服务的数量呈指数倍增加的趋势。而且,伴随着由云计算所提出来的按此付费的商业模式将使服务供应商能够根据Qos值向顾客提供不同配置的服务【3】。.因此,可以预见的是,消费者将会面对品类繁多的根据不同的QoS标准和具有不同价格的同类服务,于是一种自动选择服务的方法呼之欲出。
现在,在过去的几年里,基于QoS的网络服务及其构成方面的难题受到了极大地关注。在【4】,作者认为这种操作可以通过智能体互动法得以实现。主要的挑战就是,以一种把个体力量资本化的方式,为主题导向型和服务导向型这两种范式开发一种集成框架。本论文倡导主题导向型的网络理念(AWS)。我们提出几个关键性议题,包括合适的架构框架和其主要元素的结构(主体导向型网络服务),他们的元数据模型,辅助性技术, 集成方法,及其安装启用方法等。In [5],
在【5】,作者们展示了一种可导入网络服务选择事件的中间设备平台,可使服务的构成最大限度的满足消费者的需求,这种需求的满足通过对QoS属性的应用得以实现。同时符合使用者的限制条件和混合服务的结构的限制条件。论文对两种选择方法惊醒了描述和比较:一种是建立在本地任务水平的服务选择类型,另一种是建立在全球范围内的,整数规划任务的服务类型。
在【6】,作者们展示了一种可以提高评估效率结构,此结构还向服务商提供第三方网络服务。大部分的服务供应结构都是把主要精力集中在为遗留系统提供前端网络服务和 服务的聚集及传送等方面。这种结构主要考虑服务供应商和其商业伙伴(网络服务使这种伙伴关系成为可能)之间的静态合同。这种方法使上述结构流于死板而不能够应对商业方面的各种要求,也不能对合作伙伴的工作要求和消费者的要求做出灵活反应。而作者们所提倡的这种结构为服务商提供了一种可以灵活优化商业运作的方法。然后根据历史的运作情况,现有的环境,以及优化的商业规则,再根据消费者的需要选中和调用出来一种最佳的服务。
在【7】,作者提出一种QoS感知缺点包容的中介软件来解决这一关键难题 。我们的中介软件包括:一个使用者协调QoS模型,各种却显得包容对策,以及一种环境感知算法(这种算法可以针对有状态和无状态的网络服务的缺陷采取最佳对策)。上述中介软件的效用已经被实验证明,选择最优缺陷包容对策的算法执行性能也已被广泛的调查研究。
在【8】,作者介绍了一种针对网络服务要求的网上监督的方法这种方法包括一种符合服务要求限制的基于伙伴关系的规格,一个囊括了与顾客的要求,服务,反应,应用,资源及管理有关的五种系统事件监督模型,和一个可供不同的探测器和中介搜符合要求的事件和数据的监督框架。此框架通过打破预想限制对收集到的信息进行分析,从而达到对网络行为和网络服务进行评估的效果。
在这篇论文中,我们提倡一种QoS感知的网络服务选择法,此法可解决面对大量的网络服务时服务的构成问题。在我们所提倡的这种方法里我们运用基因算法来为消费者寻找最佳的网络服务。.为了评估我们所提倡的这种方法,我们通过实验对其加以检验。在实验中我们采用了来自世界互联网的一个真实的数据集。实验结果表明我们所提倡的方法可以极大地加快网络服务进程,优化网络服务结构。
这篇论文的结构如下:第二部分描述我们所提倡的方法即基因算法在Qos感知网络服务选择中的应用。为了检验我们的方法我们在第三部分进行了实验。最后,第四部分是我们的结论。
II. 网络服务选择中的基因算法
基因算法(以下简称GA)是建立在自然基因和选择机制上的优化算法。GA的基本概念是:一种模仿自然进化的必要的选择过程而进行设计的算法。为了把基因算法运用到具体的问题当中,就需要把问题分解成和基因相对应的原子单位。基因信息由一条比特条和比特编码解决方案集构成。这条比特条的长短不固定。于是个体们就可以建立相应的有限的基因线条,一个个体集合叫做一个群。在这里一条准则需要被确立:提个适应性函数F代表群里的每一个个体,设F(x),它的值代表了我们要解决的相关问题的个体的价值。
GA有三种运作方法:重组,选择和突变。重组的过程在这种算法里被称作交叉。也就是一个性操作过程,在此过程中分别从两条父母条里复制选中的比特,然后产出两条子条。这个选择操作可以一直凭意愿重复,直到新的一代完成。突变是由单个比特对比特条的随机性变化所引起的(可能性极小)。在这种算法里,个体被选中的几率和他们的客观的功能价值成正对应。那些作为最好的个体而被选中的个体将会被完整的带到下一代群里。这种操作被称为精英主义。
根据以上有关GA的准则,我们认为GA非常适合于为完善网络服务构成而进行网络服务选择的任务。为了在网络服务选择中运用GA,需要重点强调两点:网络服务选择的基因编码和对适应函数的定义。
1)基因编码:在基因算法中群中的一个个体代表了一个网络服务选择方案,这个方案被编入一个有n个整数的序列x1,x2…xn中,在这里n就是在工作流程中网络服务类型的总数。在基因编码计划里,每一个基因都代表了一个混合网络服务里的一个服务类别,每一个基因值则代表了这个服务类别的具体的服务。如果这个基因编码是1,那么这项网络服务就会被选择,从而成为服务结构的一部分。如果它的编码是0,这项服务就会被放弃。
2)提高适应性:在这篇论文中,为了尽可能的在现有的突变路径的基础上增加不同的突变路径,突变的很可能是染色体而不是轨迹。具体的策略如下:在对每一个染色体的突变操作之前,突变可能性用以确认染色体是否突变。如果已经突变,首先将会确认实体路径是否和当前的染色体所表现出来的路基相同。如果不同,这条实体路径将会从现有的除了当前这条线路之外的所有可见的线路中被选中。如果对象是其本身,新的染色体就会被检测确认是否和原来的染色体一样。同样的染色体会再次进行同样的操作。如果实体是一条和当前路径不同的路径,一个新的染色体就会在现有路径的基础之上创造出来。显而易见的是,检验新的染色体和老的染色体是否相同是没有必要的。
III.实验
A实验设置
在我们的评估中,我们采用了真实的QoS资料组。这个资料组是从【3,9】来的真实的QWS资料组。这个资料组包括对2500个网络服务的9种属性的测量值。表格一列出了在个资料组中这些属性和对每种属性的简要描述。对这个数据组采用了用与网络服务的商业基准工具,这些基准工具被定位于应用网络上的公共资源,它们包括UDDI注册,搜索引擎和服务门户网站等。
这个资料组的主要目标就是为网络服务搜索者提供一个基础。Eyhab Al-Masri为这个数据组搜集了5,000个网络服务,并在此数资料组上演示了个中国人养的测量值。这些服务是用我们的网络服务爬行引擎(WSCE)搜集的。大部分的网络服务都是从公共资源网络商获得的,这些公共资源网络包括:UDDI册,搜索引擎和服务门户网站。这个公共资料组由365项网络服务组成,每项服务都带有九种QWS(网络服务的质量)属性,这些属性被用来测量商业基准工具的应用。连续三天,我们对每一项服务都进行了10分钟以上的测试。关于这个资料组的更详细的资料请查阅[9-10]。
表格一
QWS资料组
QoS属性 QoS描述 单位
反应时间 发出请求和接受反应所用的时间 毫秒
有效性 成功调用|总调用的数量
生产能力 既定时间内调用的总数
成功可能性 反应的数量|请求的数量
可靠性 在所有信息中错误信息的比率
服从性 WSDL文件在多大限度上服从WSDL规格
最优方法 在多大程度上网络服务服从网络服务的互动性()
潜在因素
文件材料
B.实验结果
在这次实验中我们对所获得的结果进行了评估。.为了保证实验结果的公平性,具体的网络服务的号码是在【1,50】范围内的一个随机的整数,服务类别的号码也是在【1,10】范围内的一个随机的整数。我们首先来比较一下 结果的最优程度。字母OPT1代表了我们的方法总体使用价值。字母OPT2则代表了【11】的总体使用价值。字母OPT代表我所提倡的们的方法的最优程度,这里OPT=OPT1|OPT2。表格
II和表格III显示了结果的最优程度。
表格II
不同类别的具体的服务的最优程度
________________________________________
服务类别的号码 最优程度
________________________________________
2 98.3%
4 96.8%
6 95.8%
8 94.4%
10 93.2%
________________________________________
表格III
不同类别的服务的最优程度
________________________________________
具体服务的号码 最优程度
________________________________________
10 93.6%
20 91%
30 91.5%
40 90.3%
50 89.9%
________________________________________
不考虑网络服务在此实验中的号码,上面的结果显示我们所提倡的这种方法的最优程度要大90%以上。我们所提倡的方法在网络服务选择过程中是非常有效的。
例如从表格II可以看出我们这种方法的最优程度平均超过95%。
这意味着我们的方法在服务选择的过程中有很高的最优程度,也就是说,服务使用者可以准确的看到最适合的服务,无论其号码是多少。虽然伴随着服务类别的号码的增加,最优程度有所下降,但最优程度依然高于90%。而且,从表格III可以看出,我们的方法的平均最优程度依然高于91%。这意味着我们的方法可以在服务的选择中表现出很高的最优程度,也就是说,服务使用者可以准确地看到在众多具体服务中的最核实的服务。
从实验结果看,使用这种方法,服务使用者可以从大量的网络服务中找到符合他们的要求的最合适的网络服务。因此我们的方法对解决网络服务的结构难题是很有效地。
IV.结论
网络服务的选择是建立一种可执行的混合式网络服务。总体来说,这个选择过程包括了关于网络服务的非功能性价值的决策定制过程。由于网络服务的数量在迅速地增加,网络服务环境的QoS也是瞬息万变,因此快速选择是至关重要的。这篇论文展示了一种QoS感知的网络服务选择方法,这种方法可以解决在面对大量的网络服务时,网络服务的构成性难题。在这种方法里我们使用基因算法来找到最适合消费者的服务类型。我们使用世界互联网上的一个真实的QoS资料组来进行实验以验证我们的方法。实验结果显示我们的方法对解决网络服务的构成性难题非常有效。
参考文献
【1】选自由L. Baresi, E. Di Nitto, C. Ghezzi, an和 S. Guinea合著的《服务导向的计算机信息处理技术及应用》的第一章“具有强适应性的网络服务构成的发展规划”,第75-91页,该书出版于2000年。
【2】 Seekda!. http://webservices.seekda.com/
【3】“为以QoS为基础的网络服务构成选择天际线服务”
-----------------选自M. Alrifai, D. Skoutas, 和 T. Risse合著的《第19届关于全球网络的国际会议的会议记录》,此次会议于2000年美国南卡罗林那州首府罗利市举行。
【4】选自Y. S. Li, W. M. Shen,和H. Ghenniwa合著的《计算智能》第20章“基于主体得网络服务规划及其发展环境”678-692页,该书出版于2000年10月。
【5】选自由L. Zeng, B. Benatallah, A. H. H. Ngu, M. Dumas,
J. Kalagnanam,和H. Chang合著的《电气和电子工程师协会关于软件工程的汇报》“适合于网络服务构成的QoS感知中介软件”第30章第311-327页,发表于2000年。
【6】D. Chakraborty, S. K. Jaiswal, A. Misra, 和 A. A. Nanavati合著的
“为供应商提供评估和选择的第三方网络服务中介软件结构”于2000年在美国福罗里达州奥兰多市发表。
【7】选自由Z. Zheng 和M. R. Lyu合著的《IEEE/IFIP关于可信任的系统和网络的国际大会的会议记录:数据库2000》,,2000年六月-2000年七月,葡萄牙首都里斯本,第239-248页。
【8】选自由Q. Wang, J. Shao, F. Deng, Y. Liu, M. Li, J. Han,和 H. Mei合著的《IEEE关于计算机信息处理技术服务的汇报》的第二章“对网络服务请求的在线监督的方法”第338-2000页,发表于2000年。
【9】选自由M. Alrifai 和 T. Risse合著的《第18届关于全球网络的国际会议的会议记录》第881-890页“为了QoS感知服务构成的有效性把全球优化和地区选择结合起来”,大会于2000年在美国纽约举行。
【10】选自由E. Al-Masri 和 Q. H. Mahmoud合著的《第17届关于全球网络的国际会议的会议记录》第795-800页“对全球范围内网络服务的调查”,此次会议于2000年4月21-2000年4月25召开。此文发表于2000年。
【11】选自由D. Ardagna和B. Pernici 合著的《电气和电子工程师协会关于软件工程的汇报》第33章,第369-384页“动态过程中的适应性服务构成”,于2000年发表。
【12】选自由S. Tasaka 和 H. Yoshimi合著的《IEEE全球通信大会:全球通2000》“通过运用视频封包遗存空间性和暂时性的价值之间的权衡加强QoE在试听IP传输过程中的作用”
[/中文]
[外文]QoS-aware web Service Selection based on Genetic Algorithm in web Service
Composition Problem
Abstract—Service Oriented Architecture (SOA) has emerged as a powerful paradigm for building complex distributed web applications from simpler application components known as services. However, the increasing number of web service providers throughout the globe, have produced numerous web services providing the same or similar functionality. Then various techniques have been proposed towards selection of QoS aware web services based on the requester’s QoS requirements. In this paper, in order to select the most suitable web service among functionally similar web services, we proposed a QoS-aware web service selection approach . In this approach, we adopt genetic algorithm to select the most suitable web service
from each service class according to user’s QoS requests. Experimental results show that our proposed
approach is very effective for solving web service composition problem.
Keywords-QoS; web service; service selection; genetic
algorithm
I. INTRODUCTION
It is well known that the web is a distributed, dynamic, and large information repository. It has now evolved to encompass various information resources accessible worldwide. Then web service created. web services are selfdescribing software entities that can be advertised, located, and accessed across the Internet using a set of standards such as SOAP, WSDL, and UDDI [1]. web services encapsulate application functionalities and information sources, and promise to take cross-network application interactions one step further by enabling programmatic access to applications over the web. For a web environment, multiple web services may provide similar functionalities with different non-functional property values. Then these web services will typically be grouped together in a single community. To differentiate the members of a community during service selection, their QoS values need to be considered and it is highly recommended to be taken into account during the web service selection. It is well known that web services technology is a promising solution for addressing platform interoperability and compatibility problems faced by system integrators.
However, their adoption rate has been very slow due to the lack of QoS during the web service selection. web services technology has yet to face questions such as: “How I will know the web service will meet my performance requirements?’’ Until these questions have been solved, it is unrealistic to expect a business application in a Universal Description, Discovery and Integration (UDDI) registry. In addition, another problem becomes especially important and challenging as the number of functionally-equivalent services offered on the web at different QoS levels increases. According to [2] , there has been a more than 130% growth in the number of published web services in the period from October 2000 to October 2000. The statistics published by the web services search engine Seekda also indicate an exponential increase in the number of web services over the last three years. Moreover, it is expected that the pay-per-use business model promoted by the Cloud Computing paradigm will enable service providers to offer their (software)
services to their customers in different configurations with respect to QoS properties [3]. Therefore, it is expected that service requesters will be soon faced with a huge number of variation of the same services offered at different QoS levels and prices, and the need for an automatic service selection method will increase. At present, the problem of QoS-based web service selection and composition has received a lot of attention during the last years. In [4], the authors argue that such operations can be realized through agent-oriented interaction approaches. The key challenge is to develop an integration framework for the
two paradigms, agent- and service-oriented, in a way that capitalizes on their individual strengths. This paper proposes the notion of agent-based Web services (AWS). We address several critical issues, including the appropriate architectural framework and the structure of its main elements (agentbased Web services), their meta-model, supporting technologies, integration method, and implementation approach. In [5], the authors present a middleware platform which addresses the issue of selecting Web services for the purpose of their composition in a way that maximizes user satisfaction expressed as utility functions over QoS attributes,
while satisfying the constraints set by the user and by the structure of the composite service. Two selection approaches are described and compared: one based on local (task-level) selection of services and the other based on global allocation of tasks to services using integer programming. In [6], the authors present an
architecture to facilitate efficient evaluation and selection of 3rd party web services for service providers. Most service provider architectures have primarily focused on providing web service front ends to legacy systems, aggregating and delivering services via workflows. These architectures primarily considered static
business contracts between the service provider and its (web service enabled) business partners. This
approach makes these architectures inflexible to variations in business requirement, partners’ performance and customer requirements. Their proposed architecture provides a flexible means for service providers to optimize business performance. Then based on the historical performance, extant context, and optimizing business rules, the appropriate service is selected and invoked to serve a customer request. In [7], the author propose a QoS-aware fault tolerant middleware to attack this critical problem. Our middleware includes a user-collaborated QoS model, various fault tolerance strategies, and a context-aware algorithm in determining optimal fault tolerance strategy for both stateless and stateful Web services. The benefits of the proposed middleware are demonstrated by experiments, and the performance of the optimal fault tolerance strategy selection algorithm is investigated extensively. In [8], the author introduced an online monitoring
approach for web service requirements. It includes a patternbased specification of service constraints that correspond to service requirements, and a monitoring model that covers five kinds of system events relevant to client request, service response, application, resource, and management, and a monitoring framework in which different probes and agents collect events and data that are sensitive to requirements. The framework analyzes the collected information against the prespecified constraints, so as to evaluate the behavior and
use of Web services. In this paper, we propose a QoS-aware web service selection approach for solving web service composition problem with a great number of web services. In our proposed approach, we use genetic algorithm to find the most suitable web service for service users. In order to evaluate our proposed approach, we conduct experiments to verify our approach. In the experiment, we adopt a real QoS dataset from world internet. Experimental results indicated that our proposed approach significantly improve the web service selection process in web service composition problem. This paper is organized as follows. Section II describes our proposed method, i.e., a application of genetic algorithm in QoS-aware web service selection. In order to evaluate our approach, we take experiments in Section III. Finally, Section IV is our conclusions.
II. GENETIC ALGORITHM FOR SEVICE SELECTION
Genetic Algorithms (GA) are optimization algorithms based on natural genetics and selection mechanisms. The basic concept of GA is designed to simulate processes in natural system necessary for evolution. To apply genetic algorithms to a particular problem, it has to be decomposed in atomic units that correspond to genes. The genetic information (chromosome) is represented by a bit string and sets of bits encode the solution. The bit string may be of variable length. Then individuals can be built with correspondence to a finite string of genes, and a set of individuals is called a population. A criterion needs to be defined: a fitness function F which, for every individual among a population, gives F(x), the value which is the quality of the individual regarding the problem we want to solve. GA uses three operators: recombination, selection and
mutation. The recombination process is called crossover in this algorithm, which it is a sexual operation that creates two offspring strings from two parent strings copying selected bits from each parent. The selection operation is repeated as often as desired usually until the new generation is completed. Mutation is carried out by randomly changing the value of a single bit (with small probability) to the bit strings. In these algorithms individuals are chosen with a probability according to their objective function values. Some of these selected individuals, considered as the best individuals, are carried forward into the next generation population intact. This operation is known as elitism. According to the above descriptions about GA, we think GA is very suitable to be used to fulfill web service selection task for web service composition problem. To apply GA to web service selection, two important issues should be addressed: genetic encoding of web service selection and the definition of the fitness function.
1) Genetic encoding: An individual in the population of genetic algorithm represents a web service selection plan and it is encoded in an array of n integers x1,x2 ••• xn, where n is the total number of web service classes in the workflow of the composite web service. In the gene encoding scheme, each genetic represents a web service class in the composite web service and a value of the gene represents a concrete web services of the web service class. If the gene is 1, then the web service is selected for service compostion. If it is 0, then this service is discarded.
2) Improve Fitness: In this paper, in order to promote the probability to create different paths from the mutated path, the probability of mutation is for the chromosome instead of the locus. The concrete policy is as follows: before mutation operation of every chromosome, the probability of mutation is used to confirm whether the chromosome mutates or not. If mutation, the object path will be confirmed firstly whether it is the same as the current path expressed by thecurrent chromosome. If difference, the object path will be
selected from all available paths except the current one. If the object is itself, the new chromosome will be checked whether the new chromosome is the same as the old chromosome. The same chromosome will result in the mutation operation again. If the objects are different paths from the current path, a new
chromosome will be created on the basis of the object path. Obviously, it is not necessary to check whether the new and old chromosomes are the same.
III. EXPERIMENT
A. Expeirment Setup
In our evaluation we experimented with a real QoS dataset. The dataset is the QWS real dataset from [3, 9] This dataset includes measurements of 9 QoS attributes for 2500 real web services. Table 1 lists the QoS attributes in this dataset and gives a brief description of each attribute. The dataset was measured using commercial benchmark tools for web services, which were located using public sources on the Web,
including UDDI registries, search engines and service portals. The main goal of this dataset is to offer a basis
for Web Service researchers. For the dataset, Eyhab Al-Masri have collected 5,000 web services and performed various measurements on this dataset. The services were collected using our Web Service Crawler Engine (WSCE). The majority of Web services were obtained from public sources on the Web including Universal Description, Discovery, and Integration (UDDI) registries, search engines, and service portals. The public dataset consists of 365 Web services each with a set of nine Quality of Web Service (QWS) attributes that he have measured using commercial benchmark tools. Each service was tested over a ten-minute period for three consecutive days. For more details about this dataset we refer
the reader to [9-10].
TABLE I
QWS DATASET
QoS Attribute QoS Description Units
Response Time Time taken to send a request and
receive a response
millisecond
Availability Number of successful
invocations/total invocations
percent
Throughput Total number of invocations for
a given period of time
Invocations/sec
ond
Likelihood of
success
Number of response/number of
request messages
percent
Reliability Ratio of the number of error
messages to total messages
percent
Compliance To which extent a WSDL
document follows the WSDL
spec.
percent
Best Practices To which extent a web service
follows the Web Services
Interoperability (WS-I) Basic
Profile
percent
Latency Time the server takes to process
a given request
millisecond
Documentation Measure of documentation (i.e.
description tags) in WSDL
percent
B. Experimental Results
In this experiment, we evaluate the quality of the results obtained. The number of concrete web service is a random integer in the range [1, 50] with respect to the number of service classes ,and the number of service class is a random integer in the range [1,10] with respect to the number of concrete web services.
We first compare the optimal degree of results. The letters “OPT1” represent the overall utility value of our
approach. The letters “OPT2” represent the overall utility value of [11]. The letters “OPT” represent the optimal degree of our proposed approach, where OPT=OPT1/OPT2. Table II and Table III show the results on the optimal degree.
TABLE II
OPTIMAL DEGREE WITH DIFFERENT CONCRETE ERVICES
The number of service classes Optimal degree
2 98.3%
4 96.8%
6 95.6%
8 94.4%
10 93.2%
TABLE III
OPTIMAL DEGREE WITH DIFFERENT SERVICE CLASSES
The number of concrete services Optimal degree
10 93.6%
20 91%
30 91.5%
40 90.3%
50 89.9%
The results above show that the optimal degree of our proposed approach is larger than 90% regardless the number of web services in this experiment. This means that our approach is very effective in service selection process. For example, from Table II, the optimal degree of our approach is larger than 95% on average. This means that our approach can perform the service selection with high optimal degree, i.e., service users can accurately find the most appropriate service regardless of the number of service
classes. Although with the increasing number of service class, the optimal degree is decreasing, its optimal degree is still larger than 90%. Furthermore, from Table III, the optimal degree of our approach is larger than 91% on average. This means that our approach can perform the service selection with high optimal degree, i.e., service users can accurately find the most appropriate service from many concrete services. From the experimental results, by means of the approach, service users can find the most web service from a great
number of services according to theirs QoS requirements. Hence, our approach is effective for web service composition problem.
IV. CONCLUSTIONS
The web service selection is the process of building an executable composite web service. Generally, the selection process involves decision making in terms of non-functional properties of Web services such as QoS requirements. Since the number of web services is rapidly increasing and the QoS of the web services environment changes dynamically, the fast selection is important. This paper presents a QoS-aware web service selection approach for solving web service composition problem with a great number of web services. In this approach, we use genetic algorithm to find the most suitable web service for service users. We conduct experiments to verify our approach with a real QoS dataset from world internet. Experimental results show that our proposed approach is very effective for solving web service composition problem.
REFERENCES
[1] L. Baresi, E. Di Nitto, C. Ghezzi, and S. Guinea, "A framework for the deployment of adaptable web service compositions," Service Oriented Computing and Applications, vol. 1, pp. 75-91, 2000.
[2] Seekda!. http://webservices.seekda.com/
[3] M. Alrifai, D. Skoutas, and T. Risse, "Selecting skyline services for QoS-based web service composition," Proceedings of the 19th international conference on World wide web, Raleigh, North Carolina, USA, 2000.
[4] Y. S. Li, W. M. Shen, and H. Ghenniwa, "Agent-based Web services framework and development environment," Computational Intelligence, vol. 20, pp. 678-692, Nov 2000.
[5] L. Zeng, B. Benatallah, A. H. H. Ngu, M. Dumas, J. Kalagnanam, and H. Chang, "QoS-aware middleware for Web services composition," IEEE Transactions on Software Engineering, vol. 30, pp. 311-327, 2000.
[6] D. Chakraborty, S. K. Jaiswal, A. Misra, and A. A. Nanavati, "Middleware architecture for evaluation and selection of 3rd party web services for service providers," Orlando, FL, United states, 2000, pp. 647-654.
[7] Z. Zheng and M. R. Lyu, "A QoS-aware fault tolerant middleware for dependable service composition," Proceedings of IEEE/IFIP International Conference on Dependable Systems and Networks, DSN 2000, June 29, 2000 - July 2, 2000, Lisbon, Portugal, 2000, pp. 239-248.
[8] Q. Wang, J. Shao, F. Deng, Y. Liu, M. Li, J. Han, and H. Mei, "An online monitoring approach for Web service requirements," IEEE Transactions on Services Computing, vol. 2, pp. 338-351, 2000.
[9] M. Alrifai and T. Risse, "Combining global optimization with local selection for efficient QoS-aware service composition," Proceedings of the 18th international conference on World Wide Web, New York, United States:, 2000, pp. 881-890.
[10] E. Al-Masri and Q. H. Mahmoud, "Investigating web services on the world wide web," Proceedings fo the 17th International Conference on World Wide Web, April 21, 2000 - April 25, 2000, Beijing, China, 2000, pp. 795-800.
[11] D. Ardagna and B. Pernici, "Adaptive service composition in flexible processes," IEEE Transactions on Software Engineering, vol. 33, pp. 369-384, 2000.
[12] S. Tasaka and H. Yoshimi, "Enhancement of QoE in Audio-Video IP Transmission by Utilizing Tradeoff between Spatial and Temporal Quality for Video Packet Loss," Proceedings of IEEE Global Telecommunications Conference, GLOBECOM 2000.[/外文]
|