Semantic Services[MSOffice1]

WWRF WG2 White Paper

WG2-WPxx

Version 0.431

Josef Noll1,Matthias Wagner2,Erik Lillevold1,Anna V. Zhdanova3, Kashif Iqbal4

1 University Graduate Center, UniK, Kjeller, Norway

2 DoCoMo, EuroLab, Germany

3 CCSR, University of Surrey, UK

4DERI, National University of Ireland, Galway, Ireland

Table of Content

1.Introduction

2.Objective and Scope

2.1Objectives

2.2Scope

2.3Approach

2.4Semantic Service Scenarios

3.From Web Services towards Semantic Services

3.1Evolution to future service provisioning in Business Enterprises

3.2Semantic Web

3.2.1 Semantic support technologies

3.2.2 Mobile Ontology: Ontology Adoption for Mobile Services

3.3Semantic Web Services

3.3.1 OWL-S (already explained in section 3.2.1)

3.3.2 WSMF (WSMO, WSML, WSMX)

3.3.3 WSDL-S

3.3.4 SWSF (SWSO, SWSL)

3.3.5 SAWSDL

4.Challenges

4.1Service Realisation and Expectations

4.2Open issues

4.2.1 Semantic service architecture interfaces

4.2.2 Mediation of ontologies

4.2.3 Decision making (reasoning, rules)

4.3Mobile Semantic Services

4.3.1 Mobile Service Provision

4.3.2 Semantic Web services vs. Mobile services

5.Business Considerations

5.1Tools and Practical Usage of Semantics in Web Services

5.2Business evaluation

6.Related Work

7.Conclusions

8.Acknowledgements

9.References

1.Introduction...... 3

2.Objective and Scope...... 3

2.1Objectives...... 3

2.2Scope...... 3

2.3Approach...... 5

2.4Semantic Service Scenarios...... 5

3.From Web Services towards Semantic Services...... 6

3.1Evolution to future service provisioning in Business Enterprises...... 6

3.2Semantic Web...... 7

3.2.1 Semantic support technologies...... 7

3.2.2 Mobile Ontology: Ontology Adoption for Mobile Services...... 15

3.3Semantic Web Services...... 20

4.Challenges...... 20

4.1Service Realisation and Expectations...... 21

4.2Open issues...... 21

4.2.1 Semantic service architecture interfaces...... 21

4.2.2 Mediation of ontologies...... 21

4.2.3 Decision making (reasoning, rules)...... 21

4.3Mobile Semantic Services...... 21

4.3.1 Mobile Service Provision...... 21

4.3.2 Semantic Web services vs. Mobile services...... 21

5.Business Considerations...... 21

5.1Tools and Practical Usage of Semantics in Web Services...... 21

5.2Business evaluation...... 22

6.Related Work...... 22

7.Conclusions...... 22

8.Acknowledgements...... 22

9.References...... 22

1.Introduction...... 4

2.Objective and Scope...... 4

2.1Objectives...... 4

2.2Scope...... 4

2.3Approach...... 6

2.4Semantic Service Scenarios...... 6

3.From Web Services towards Semantic Services...... 7

3.1Evolution to future service provisioning in Business Enterprises...... 7

3.2Semantic Web...... 8

3.2.1 Semantic support technologies...... 8

3.3Semantic Web Services...... 16

4.Challenges...... 17

4.1Service Realisation and Expectations...... 17

4.2Open issues...... 17

4.2.1 Semantic service architecture interfaces...... 17

4.2.2 Mediation of ontologies...... 17

4.2.3 Decision making (reasoning, rules)...... 17

4.3Mobile Semantic Services...... 17

4.3.1 Mobile Service Provision...... 17

4.3.2 Semantic Web services vs. Mobile services...... 18

5.Business Considerations...... 18

5.1Tools and Practical Usage of Semantics in Web Services...... 18

5.2Business evaluation...... 18

6.Related Work...... 18

7.Conclusions...... 18

8.References...... 18

History:

  • v0.3, first texts from SPICE (UniS) inserted
  • v0.2 Text in the current document is just copied, needs to be adapted/edited
  • v0.1 only TOC

1.Introduction

The next generation (beyond 3G) of mobile services and applications must offer ever increasing levels of value and differentiation (i.e. they have to be attractive, intuitive and easy-to-use, with personalization and ubiquitous access.). These services must also be developed easily, deployed quickly and, if necessary, altered efficiently.

2.Objective and Scope

(copy from previous WhitePaper7) with WWRF WG2 architecture– need to be edited

2.1Objectives

The objective of the WWRF WG2 is to create a technology independent blue print of next generation service platform architecture as a basis for a discussion with the different players; this includes the understanding why market participants have come up with new architectures as well as a clear understanding about the underlying network architectures.

In this context it is intended to elaborate the unique features to be provided in the corresponding service domains by telecommunication and IT industries. Also, the requirements of 3rd party providers have to be considered seriously. Of course, the impact on today’s’ decisions can not be excluded always guided by the availability of technologies.

For WWRF WG2 these objectives intend to bring the WG2 results, which are on a pure functional level so far, to a more system architecture level, identifying (high level) components, their distribution and their interworking (interfaces and protocols) based on intensive discussions among experts from operators, manufacturers, content providers, IT/software industry, and academia.

2.2Scope

A next generation mobile service network has to establish its own value proposition between the stakeholders of next generation mobile service provisioning. It has to place its unique selling points into the overall service centric environment dealing with a number of different access systems. Mobile networks have the potential to act in a central role within this service environment and is therefore required to be capable to act as a service control environment. It is not the intension to copy all successful internet services but to support them in an efficient and trustworthy way.

The service platform architecture is interfacing to the service and applications at the upper layer of a communication system. The Service Platform Architecturecovers service support components (such as Generic Service Elements), their relationship and their internal and external interfaces. It usually targets the upper system layers as depicted in Figure 1, however some functions or parts of them are residing in lower system layers, e.g., location or mobility. When considering ubiquitous communications environments such as sensor networks the layering is diminishing anyway.

Figure 1: Target area of the WG2 architecture work shown in the reference model

Service Architecture for future mobile systems must be able to satisfy requirements from a broad range of services and different kind of service usage. This can be reflected in the consideration and analysis of various future life scenarios. With this approach a scenario centred design process can be applied.

During requirement analysis, a number of different scenarios have been considered. They have been obtained from a number of different sources, including project results, dedicated studies and published literature. The following list provides their references:

  • ISTAG; Scenarios for Ambient Intelligence in 2010; Final Report, Feb 2001, EC 2001. Available at
  • The Activities of 4th Generation Mobile Communications Committee (FLYING CARPET Ver.2.00), March 2004, available at
  • R. Tafazolli (Editor), Technologies for the Wireless Future: Wireless World Research Forum (WWRF), ISBN 0-470-01235-8, Wiley 2005
  • ETSI/TIA Project mesa scenarios, Fall 2003. Available at
  • MIT Oxygen project scenarios: Pervasive Human-centred computing. Available at
  • NTT DoCoMo Vision 2010 scenarios. Multimedia content available at
  • CyPhone Mediaphone Project ­ Taxi Trip Scenario, 1999. Available at
  • IST-SIMPLICITY project scenarios. Available at
  • The Vision of ITEA (SOFTEC Project); Technology Roadmap on Software Intensive Systems. Available at
  • Bo Karlson et al, Wireless Foresight: Scenarios of the Mobile World in 2015, Wiley, 2003

Although there are a lot of scenarios available, their impact on service architectures is rarely obvious. The reason might be that service architectures have not been in the focus of the scenario creator at all.Alternatively, the scenarios might be tailored to specific service architectures, i.e. those activities do not comply with the scenario centred design process. In cases where scenarios were in fact considered for the design process, these are often describing very specific aspects and the resulting service architectures are very much tailored to them.

To gain the desired demands from future life scenarios on service architectures it is required to evaluate a selection of scenariosto derive commonalities and their specific features. A first analysis resulted in a number of aspects which can be taken as initial input for the service architecture design:Future Services will be context-sensitive, adaptive and personalized. They will be available in different networks, with different bandwidth/QoS and for different devices or multimodal UIs respectively. Service architectures need to support sophisticated charging and billing, security and privacy, identity management, DRM, and trust. The complete service lifecycle is to be reflected: from service creation and composition to service discovery and delivery.

2.3Approach

example, needs update

WWRF WG2 proposes the following approach to address the above stated objectives and scope of future service architectures in order to avoid unsuccessful investments:

  • Identify unique selling points as Value Added Services (= high level features) provided by Mobile Service Platforms
  • Verify the Value Added Services with service scenarios
  • Identify a suitable architecture supporting those features and addressing the above stated scope
  • Design of an architecture blueprint in terms of high level components and interfaces
  • Identify platforms and software engineering methods for its realization

2.4Semantic Service Scenarios

  • mobile traveller
  • emergency case
  • spice project?

3.From Web Services towards Semantic[EL2] Services

3.1Evolution to future service provisioning in Business Enterprises

Personalization is one of the most important characteristics of future telecommunications. Personalization means tailoring of services and applications to the very specific needs of a user to a ubiquitous comfortable service environment together with a single bill service independent of the partners involved in the end to end value chain. Personalization also means to allow fully manual service configuration, carrier access without any automatic vertical handover procedures and individual invoicing of the individual service partners, in case this is required by any user for any reasons.

Market participants are faced the following trends:

  • Content and services are distributed across domains including fixed line and mobile networks as well as open (e.g., world wide web) and closed service environments. Business relations might be distributed.
  • Internet services are mainly used by fixed line access, whereas hotspots just start to
  • Technology evolution separates call control and transport. The directions of centralize or spread out of the service control are regarded as a chance and a threat for mobile operators.
  • Transport charges are dropping tremendously due to excessive IP transport capacities in back-bone networks and due to increased competition in access networks (flat rates).
  • There is a strong impact of IT technologies on telecommunications protocols, system design and interfaces.
  • A growing importance of peer to peer services integrates the single user into the group of service and application providers.
  • An increasing importance of ubiquitous computing technologies (sensor networks, RFID, NFC) leads to pervasive interaction with service systems and provide new possibilities for services based on rich information.
  • Trust in context aware services still suffers from privacy concerns of the users. In addition legal aspects do also restrict provides in offering rich context aware services.

These aspects lead to following conclusions:

  • Users will maintain several business relations in different service domains.
  • Access procedures including authentication and authorization will become more and more complex and will happen more often.
  • IP based services will dominate.
  • Service control is not per definition in the domain of the telecommunication operator.
  • Telecommunication network operators can hardly survive by providing bit pipes only. Value added services are a prerequisite for future business of telecommunication operators. Mobile network operators have to identify their unique selling points and to force them into the market.
  • A trustworthy position in the market is important for the acceptance by the users.

Figure 2 illustrates our vision of a multi domain network and service environment. Mobile services platforms of telecommunication providers and mobile operators in parallel to 3rd party service environments provide service through application servers to the customers. Additional functionalities, e.g., identity manager or billing, are depicted as separate functions that may be provided also be 3rd parties.

3.2Semantic Web

  • The Semantic Web Stack
  • Ontologies
  • Tools and practical usage of Semantics in Web services

3.2.1Semantic support technologies

taken from previous white paper, need to be edited

In order to enable the meaningful communication between different GSEs, developed by different vendors, a common understanding of used terms and definitions need to be achieved. But as there is no global common understanding and use of terms this understanding needs to be established between communicating parties on the fly. In order to enable this mechanism, semantic descriptions, using taxonomies and ontologies need to be used. The following section will introduce the state of the art in respect to ontologies and taxonomies, describing GSEs, and will

Ontologies

Ontologies are for knowledge sharingandreuse, while languages such as XML are perfect to express information in a structured and efficient way. To be able to discuss with one another, communicating parties need to share a common terminology and meaning of the terms used. Otherwise, profitable communication is infeasible because of lack of shared understanding. With software systems, this is especially true—two applications cannot interact with each other without common understanding of terms used in the communication. Until now, this common understanding has been achieved awkwardly by hard-coding this information into applications. This is where ontologies come into the picture. Ontologies describe the concepts and their relationships—with different levels of formality—in a domain of discourse. An ontology is more than just a taxonomy (classification of terms) since it can include richer relationships between defined terms. For some applications, a taxonomy can be enough, but without rich relationships between terms it is not possible to express domain-specific knowledge except by defining new terms.

Ontologies have been an active research area for a long time. The hardest issue in developing ontologies is the actual conceptualisation of the domain. Additionally, to be shared, the ontologies need a representation language. Languages like XML that define structure of a document, but lacks semantic model, are not enough for describing ontologies—intuitively an XML document may be clear, but computers lack the intuition. In recent years ontology languages based on Web technologies have been introduced. DAML+OIL [2], which is based on RDF Schema [3], is one such language. It provides a basic infrastructure that allows machines to make simple inferences. Recently, DAML+OIL language was adopted by W3C, which is developing a Web Ontology Language (OWL) [4] based on DAML+OIL. Like DAML+OIL, OWL is based on RDF Schema [3], but both of these languages provide additional vocabulary—for example relations between classes, cardinality, equality, richer typing of properties, characteristics of properties, and enumerated classes—along with a formal semantic to facilitate greater machine readability. The OWL language has a quite strong industry support, and therefore it is expected to become a dominant ontology language for Semantic Web. In this paper, we use OWL language when giving examples of ontology encoding.

The following sections tackle different aspects associated with ontologies. They will provide in particular a short description of the OWL language, the OWL-S ontology and will then point out critical issues to be deal with as far as ontologies construction, usage and management & maintenance are concerned.

Ontology construction

This section gives an introduction to OWL and OWL-S and an overview of the Protégé tool which provide an ontology editors and tools to access information available within the ontology.

OWL – The Web Ontology Language

OWL - the Web Ontology Language - is intended to provide the explicit meaning (semantics) of information while languages such as XML mainly focus on the information structuring.

OWL then facilitates the automatic handling and understanding of information by computer program (for instance reasoning about this information) and then enhances the level of cooperation between machines. OWL goes far beyond previous initiatives such as RDF, RDF-S (RDF schema) providing extended capabilities and expressive power. OWL has been built on top of W3C recommendation stack XML –> XML-Schema -> RDF -> RDF-Schema. With respect to these languages, OWL adds more vocabulary for describing the classes and the properties that holds between classes and individuals. This enhanced vocabulary includes set-theoretic constructs such as cardinality, intersection, characteristics of properties and enumerated classes. As far as expressive power is concerned, OWL is declined into three dialects (from the less expressive but most decidable and complete) to the most expressive (but less decidable and complete) that are respectively OWL-lite, OWL-DL and OWL-FULL.

Very quickly and in addition to the RDFS construct, OWL-lite provides equivalence between classes and properties, equality and inequality between individuals (member of a class), inverse of a property. Properties can be stated as transitive and symmetrical and restrictions can be applied to cardinality and to values on properties associated to classes. OWL-DL and OWL-FULL use exactly the same vocabulary (extending OWL-LITE) but some restrictions apply to OWL-DL (making the stuff complete and decidable). They provide enumerated classes, classes disjunction, union, complement and intersection and finally Cardinalities are no more restricted to 1 or 0. Please refer to OWL Web Ontology Language guide for more details about the specifications of these three dialects and full details about their capabilities.

OWL-S – Semantic Mark-up Language for Web-services[K3]

OWL-S has been especially designed to allow the definition of (web-) service ontologies. It is based on the OWL language briefly described above. OWL-S is expected to allow computer software or agent to make intelligent use of web-services and will, non-surprisingly focus on

  • service formal descriptions (aka service profile): or in other words: “What the service does”. Useful for advertising, registering and discovering services;
  • process model: providing a detailed description of the service operations
  • grounding: which provides the necessary details to communicate with the service via exchange of messages.

These three aspects can be summarised in the following picture.

Figure 1: the three concepts of OWL-S

This Web Service ontology will then propose a main class service. With this class are attached three properties (as described above) the name of which are presents, describedBy and supports. Their respective range (as described in OWL) are respectively the following classes: serviceProfile, serviceModel and serviceGrounding.

Obviously the ontology defines only the framework – that is the vocabulary – needed to describe the three aspects of a (web-) service. The exact content of these classes will vary from a given service to another. The three following sections introduce in more detail the purpose of the three “aspects” of a service description, which can also be used for describing GSE functionalities: