service oriented middleware

J Internet Serv Appl (2011) 2:23–45 DOI 10.1007/s13174-011-0021-3 O R I G I NA L PA P E R Service-oriented middleware for the Future Internet: state of the art and research directions Valérie Issarny · Nikolaos Georgantas · Sara Hachem · Apostolos Zarras · Panos Vassiliadist · Marco Autili · Marco Aurélio Gerosa · Amira Ben Hamida Received: 17 February 2011 / Accepted: 28 April 2011 / Published online: 25 May 2011 © The Brazilian Computer Society 2011 Abstract Service-oriented computing is now acknowl- edged as a central paradigm for Internet computing, sup- ported by tremendous research and technology development over the last 10 years. However, the evolution of the In- ternet, and in particular, the latest Future Internet vision, challenges the paradigm. Indeed, service-oriented comput- ing has to face the ultra large scale and heterogeneity of the Future Internet, which are orders of magnitude higher than those of today’s service-oriented systems. This arti- cle aims at contributing to this objective by identifying the V. Issarny (�) · N. Georgantas · S. Hachem INRIA, Le Chesnay, France e-mail: valerie.issarny@inria.fr N. Georgantas e-mail: nikolaos.georgantas@inria.fr S. Hachem e-mail: sara.hachem@inria.fr A. Zarras · P. Vassiliadist University of Ioannina, Ioannina, Greece A. Zarras e-mail: zarras@cs.uoi.gr P. Vassiliadist e-mail: pvassil@cs.uoi.gr M. Autili Università degli Studi di L’Aquila, L’Aquila, Italy e-mail: marco.autili@di.univaq.it M.A. Gerosa University of São Paulo (USP), São Paulo, Brazil e-mail: gerosa@ime.usp.br A.B. Hamida PetalsLink, Toulouse, France e-mail: amira.ben-hamida@petalslink.com key research directions to be followed in light of the lat- est state of the art. This article more specifically focuses on research challenges for service-oriented middleware de- sign, therefore, investigating service description, discovery, access, and composition in the Future Internet of services. Key word word文档格式规范word作业纸小票打印word模板word简历模板免费word简历 s Future Internet · Service-oriented computing · Service-oriented middleware 1 Introduction Service-Oriented Computing (SOC) is now largely accepted as a well-founded reference paradigm for Internet com- puting [102]. Under SOC, networked devices and their hosted applications are abstracted as autonomous loosely coupled services within a network of interacting service providers, consumers (aka clients) and registries according to the service-oriented interaction pattern (see Fig. 1). Still, despite the remarkable progress of the SOC para- digm and supporting technologies in the last 10 years, sub- stantial challenges have been set through the evolution of the Internet. Over the years, the Internet has become the most important networking infrastructure, enabling all to create, contribute, share, use, and integrate information and knowl- edge by all. As a result, the Internet is changing at a fast pace and is called to evolve into the Future Internet, i.e., a feder- ation of service- and self-aware networks that provide built- in and integrated capabilities such as: service support, con- textualization, mobility, security, reliability, robustness, and self-management of communication resources and services [133]. Practically, the Future Internet vision challenges all the SOC architectural layers, from the bottom to the top: ser- vice foundations as formed by the service-oriented middle- 24 J Internet Serv Appl (2011) 2:23–45 Table 1 The Future Internet constituents Constituent Definition Reference Internet of content Content is any type and volume of media. Content may be prerecorded, cached or live, static or dynamic, monolithic or modular. Content may be combined, mixed or aggregated to generate new content and media. It may vary from a few bits (e.g., the temperature that a sensor has measured) to interactive multi-media sessions and immersive complex and multidimensional virtual/real worlds’ representations. [48] Internet of services An umbrella term to describe several interacting phenomena that will shape the future of how services are provided and operated on the Internet. The Internet of Services also comprises the various sets of Internet Applications including pervasive/immersive/ambient, industrial/manufacturing, vehicular/logistics, financial/ePayment/eBusiness, power network control/eEnergy, eHealth, and eGovernment applications. [101] Internet of things A global network infrastructure, linking physical and virtual objects through the exploitation of data capture and communication capabilities. This infrastructure includes existing and evolving Internet and network developments. It will offer specific object-identification, sensor and connection capability as the basis for the development of independent cooperative services and applications. These will be characterized by a high degree of autonomous data capture, event transfer, network connectivity and interoperability. [33] Fig. 1 Service-oriented interaction pattern ware realizing the runtime infrastructure, service composi- tion, and service management and monitoring [102]. In this context, the goal of this article is to highlight research di- rections in the area of service-oriented computing in the Fu- ture Internet, based on today’s state of the art. However, due to the breadth of the area, the article focuses more specifi- cally on the study of the challenges posed to the middleware layer. Briefly stated, Service-Oriented Middleware (SOM) supports the service-oriented interaction pattern through the provision of proper functionalities for deploying, publish- ing/discovering and accessing services at runtime. SOM commonly also provides support to realize more complex composite services by integrating simpler ones, where it should be acknowledged that this contributes to the upper service composition layer. In accordance with the above, this article starts by set- ting the overall challenges and requirements posed by the Future Internet, which in particular relates to its expected ultra large scale, heterogeneity, and mobility. The article is then structured in relation to the essential functionalities of Service-Oriented Middleware, i.e., service description, ac- cess, discovery and composition, surveying related state of the art and Future Internet challenges for each one of them. Precisely, in Sect. 2, we provide our definition of the Future Internet vision and major challenges that come along with it. Then, in Sect. 3, we survey the description of services that needs to be provided for enabling effective service use in the greatly complex Future Internet environment. In Sect. 4, we concentrate on service discovery in the Future Internet, with a special focus on the organization, management and distribution of supporting service registries. In Sect. 5, we study middleware support for service access, where we high- light the key role now played by the Enterprise Service Bus paradigm as well as the evolution needed to meet Future In- ternet requirements. In Sect. 6, we focus on decentralized choreography-based composition in the Future Internet, and associated modeling and runtime support. Finally, the con- clusions are presented in Sect. 7. 2 Future internet challenges and requirements The Future Internet has become the main focus of several research and development initiatives all over the world, in- cluding initiatives in the EU,1 USA,2 China,3 Korea,4 and Japan.5 However, despite the great interest in the Future In- ternet, no common definition of it has been adopted yet. Still, considering that the Future Internet will result from the evolution of today’s Internet, the Future Internet can be defined as the union and cooperation of the Internet of Con- tent, Internet of Services, and Internet of Things, supported by an expanding network infrastructure foundation. Those core domains, elements of which we find already in today’s Internet, are not fully established yet and will emerge with the foreseen evolution of services, content, objects and net- works, as summarized in Table 1. 1http://www.future-internet.eu. 2http://www.nets-find.net. 3http://www.cstnet.net.cn/english/cngi/cngi.htm. 4http://fif.kr. 5http://akari-project.nict.go.jp/eng/overview.htm. J Internet Serv Appl (2011) 2:23–45 25 Table 2 The Future Internet challenges Challenges Today’s internet Toward the future internet Scalability 1 billion Personal Computers (2008)(*), 647 million smartphones (2010) [44] 1.78 billion Personal Computers (2013), 1.82 billion smartphones (2013)(+) 5 exabytes of data (2005) [52] 990 exabytes of data (end of 2012) [143] 104 services (2007) [4] Billions of services [133] 10 billion terminals (2010) [2] 100 billion terminals (2015) [2] Consumer Internet traffic of 12.684 exabytes/month (2010) [43] Consumer Internet traffic of 42.070 exabytes/month (2014) [43] Heterogeneity Islands of interconnected objects Internet-scale connection of highly heterogeneous objects (vehicles, sensors, mobiles devices, home appliances, etc.) [15] Emergence of heterogeneous services provided on the Cloud such as Software as a Service (e.g., Google apps) or Infrastructure as a Service (e.g., Storage services at Amazon) [149] Cloud Computing enabling to provide everything as services, spanning different business and technical domains Service/content mashups leading to the provision of new, diverse services by prosumers Global-scale services/content mashups creating new services/content with different types and formats Mobility Mostly (mobile) IPv4, which suffers from scalability issues etc.; even IPv6 has issues in mobile situations (e.g., due to the use of home agents/addresses) [101] Global-scale mobile Internet that requires revisiting communication/routing solutions [4] Wide-spread usage of smart mobile devices with limited resources (2 billion users) Global scale usage of smarter mobile devices with ever-growing resource needs Awareness & Adaptability Ad hoc solutions to network, content & service adaptation Large scale content sharing, service provisioning, mobile connectivity that require autonomic adaptation and therefore awareness of content, networks and services [101] Security, Privacy & Trust Safety and security requirements still an issue for today’s Internet Integrating real world objects, more users, more information, more services in the Internet intensifies the necessity for safety and security solutions (*)http://www.gartner.com/it/page.jsp?id=703807 (+)http://www.gartner.com/it/page.jsp?id=1278413 In general, the Future Internet is setting significant chal- lenges over the computing and networking environments, as it magnifies the features of the already challenging Internet of today (see Table 2). Specifically, key challenges posed by the Future Internet relate to and are amplified by the highly correlated nature of the following requirements: • Scalable internet: The Internets of Content, Services, and Things are confronted with scalability issues due to the increasing number, size, and quality of their networked entities, which is further exacerbated by the empower- ment of users who are now becoming “prosumers” [101, 105, 128]. For instance, simply considering the Internet of Things, the large amount of new information available through things needs to be comprehensively managed and aggregated to provide useful services [101]. • Interoperable internet: The Future Internet will be hetero- geneous in many dimensions, related to physical objects, networks, services and data, which presents a significant challenge for sustaining the Future Internet vision [101]. In particular, appropriate semantic technologies, shared standards and mediation are required to assure interoper- ability of heterogeneous entities such as things, sensors, and networks [132]. • Mobile internet: Unlike the current Internet, mobility should be natively integrated in the design of the Future Internet. Indeed, an essential challenge for the Future In- ternet lies in the explicit design of a protocol for a mobile wireless world given that the majority of the connected entities are now mobile. • Aware and adaptive internet: Awareness and related adaptability are common requirements for sustaining the Future Internet, be it at the service, content or physical object level. Issues to be addressed include: adapting the Web by and for users, adapting the network to shared me- dia and vice versa, providing personalized content and 26 J Internet Serv Appl (2011) 2:23–45 Fig. 2 Service-oriented computing in the Future Internet media to users, providing context-aware and personalized dynamic services [101, 128, 132]. • Safe internet: Trust, privacy and security are sensitive cross-domain issues that the current Internet is facing and remain critical challenges for the Future Internet. With the global-scale communications and exchange of informa- tion, users’ mobility and the limited resources their de- vices may have, as well as the Future Internet’s “aware- ness” of users, their data, and their surroundings, it be- comes crucial to find appropriate solutions that will pro- tect users. Indeed, current security mechanisms are unfit in such an open, dynamic, and aware setting. The following sections point out research directions for service-oriented middleware in light of the latest state of the art and the above requirements posed by the Future Internet. The remainder specifically concentrates on the challenges that arise for the base functionalities of service-oriented middleware in the Future Internet (see Fig. 2), i.e., service description, discovery, access, and composition. 3 Service description Service description is a fundamental element in SOC, as it determines the information that a service needs to expose to its environment for enabling its unambiguous identifica- tion and use. All other information internal to the service is simply out of the scope of SOC. 3.1 State of the art Information included in service description varies depend- ing on the complexity and the intended use of the service. Accordingly, a number of service description languages Fig. 3 Service description have been proposed to cover different description aspects and are currently in use, some of them having reached the status of standard and other still being the subject of re- search. Most of the related initiatives focus on Web Ser- vices6 being the dominant technology for SOC. W3C and OASIS are the two leading standardization bodies in this area. Besides Web Services, the Semantic Web initiative7 and related technologies have produced a significant change in the way services are perceived and described. In particu- lar, the innovation is to make explicit the business or user- domain semantics of services, so far implied by the syntax of their descriptions. Employing syntactic descriptions either makes service semantics ambiguous or calls for a syntax- level agreement between service developers/providers and service users, which is too restrictive for the inherent loosely coupled character of SOC. Service semantics are made ex- plicit by reference to a structured vocabulary of terms (on- tology) representing a specific area of knowledge. Ontology languages support formal description and machine reason- ing upon ontologies; the Web Ontology Language (OWL)8 6http://www.w3.org/2002/ws/. 7http://www.w3.org/2001/sw/. 8http://www.w3.org/2004/OWL/. J Internet Serv Appl (2011) 2:23–45 27 is the standard established by W3C. In the following, we dis- cuss the most common elements of service description (see Fig. 3) and survey their coverage by the most widely used service description languages. Service profile provides a high-level business description of a service, which may include both human-oriented in- formation (e.g., what the service does and service provider information) and machine-oriented elements. The latter, in particular, may range from a simple service name to a pre- cise semantic characterization of the service comprising its provided high-level functionalities as well as its high-level Inputs, Outputs, Preconditions, and Effects; these are col- lectively denoted as IOPEs. Inputs specify the data required by the service for its execution and Outputs specify the data provided by the service as result of its execution. In addition, Preconditions need to be fulfilled before the service may ex- ecute, while Effects specify the impact of the service on the state of the world besides its Outputs. The notion of Service Profile was created or at least made popular by OWL-S. OWL-S9 is an OWL ontology for de- scribing Web services. In OWL-S, a service description is composed of three parts: the service profile is complemented by the process model and the service grounding (see related paragraphs below). The service profile provides semantic descriptions of the service’s capabilities (the OWL-S term for service’s high-level functionalities) in terms of IOPEs. OWL-S was a candidate for becoming the W3C standard for semantic service description; however, the winner was SAWSDL, discussed in the paragraph on Service Interface below. The expressive power of Service Profile has been widely acknowledged in semantic service matching approaches [100]: matching between a requested and a provided service profile is typically the first step in service discovery and selection (see Sect. 4). However, while matching between requested and provided inputs and outputs is commonly ap- plied, there is much less use of preconditions and effects. Besides, there is less agreement within the SOC community regarding how PEs should be specified and used. After a number of submissions to the W3C of candidate rule lan- guages for the Semantic Web (suitable for being used in PEs specification and reasoning), the outcome was the creation of a W3C working group studying a Rule Interchange For- mat,10 a future standard for exchanging rules among rule systems, as no single one-fits-all rule language could be identified. Service interface specifies the set of observable lower- level (with respect to the functionalities in the Service Pro- 9http://www.w3.org/Submission/OWL-S/. 10http://www.w3.org/TR/rif-overview/. file) atomic operations that a service can perform in coor- dination with its environment, along with their input/output parameters. Service Interface is the fundamental and manda- tory element of service description, as it technically enables the access to a service as a software component (see Sect. 5). Web Services Description Language (WSDL), now in its 2.0 version,11 is traditionally the language for describing Web service interfaces, and a distinctive element of the Web Ser- vices technology. SAWSDL12 is the W3C Recommenda- tion for adding semantic annotations to WSDL and XML Schema. Such annotations can be expressed in any ontology language, most often in OWL. Annotations can be added to WSDL interfaces, operations, and the XML Schema types of their input/output parameters. Moreover, SAWSDL sup- ports the introduction of two-way transformation mappings between XML Schema types and corresponding semantic concepts. This enables the interoperability between syntacti- cally mismatching input/output parameters upon service in- vocation. In the case of OWL-S, a service interface is spec- ified semantically in the process model and syntactically in the service grounding, with appropriate mapping between the two. WSDL is again used in OWL-S service grounding. In parallel to the development of WS-∗ technologies for Web Services, REpresentational State Transfer (REST) [55] was introduced as an architectural style and an alternative way (or a return to fundamentals) for enabling services on the Web (RESTful Web services) by using the standard Web mechanism: any entity on the Web is a resource at some URI and can be accessed with the standard HTTP operations. The advantages of REST are its universality and the uniform service