UNIVERSAL, ONTOLOGY-BASED DESCRIPTION OF DEVICES

AND SERVICES

Mihail Bratoev and Boyan Bontchev

Department of Software Engineering, Faculty of Mathematics and Informatics

Sofia University5, James Bourchier blvd., 1164 Sofia, Bulgaria

Keywords: Device ontology, Service description, OWL-S, Service discovery.

Abstract: A great problem with existing service discovery middleware solutions (e.g. UPnP, Bluetooth, and Jini etc)

is that services available in one platform cannot be easily accessed from services based on another platform

which is known as “hidden service problem”. One of the reasons for this incompatibility is a different

method used for service description in each of these platforms. This paper highlights the importance of

unified, platform-neutral, semantic-oriented device and service descriptions to achieve seamless integration

between different technologies and systems for advertising, discovering and invoking services. Moreover, it

presents an ontology-driven approach for describing devices and their services which will enable service-

seeking peers to reason about available services and devices, and make intelligent and informed decisions

regarding which services to use, and how. We argue that having separate descriptions (ontologies) for

device and services that this device offers aims to be more flexible, modular and therefore better design. The

proposed device ontology is to be used for description of characteristics of the device that offers a specific

service. For development of semantic models of services our analysis shows that making several

enhancements of OWL-S (using OWL subclassing) will enable its use as common language for service

description. Finally we outline some of the future work that gives a real perspective of the proposal.

1 INTRODUCTION

The widespread deployment of inexpensive

communications technology and network-enabled

end devices poses an interesting problem for end

users: how to locate a particular network service or

device out of hundreds of thousands of accessible

services and devices. Service offers have to be

explicitly described and published, whereas service

requests need to be paraphrased and compared with

the offer descriptions. For this reason, what is

needed is a dedicated service description language.

Service oriented approach to design and

implement modular, loosely-coupled distributed

software architectures is supported by special

middleware technologies for service advertisement,

discovery and control. The main problem with the

existing service discovery technologies is that as a

whole, these technologies remain incompatible each

other. That leads to the so called isolated “islands”

of interoperability (Nakazawa J., 2006), where it is

possible one service to be available in general but

not to be discoverable by the mechanism used by its

potential clients. Since different discovery systems

use different formats for service description, we

believe that providing a mean for automated

translation between different service description

formats (languages) used by these technologies is an

important step towards seamless integration and

mutual understanding across platforms. Such a mean

should possess good abilities for semantic

expressivity required for an adequate representation

of requirements and capabilities of system services.

Interoperability at semantic level will significantly

improve service reuse and discovery and, as well,

strongly assist composition of new services and

enable business process integration. Generally

speaking, there are two possible ways for translation

of service descriptions – direct and mediated

(Nakazawa J., 2005). Mediated translation uses a

common intermediate language for protocol

independent description of services and is

considered more practical approach where a lot of

target language is available. The main disadvantage

hete – a possible loss of semantics during translation

105

Bratoev M. and Bontchev B.

UNIVERSAL, ONTOLOGY-BASED DESCRIPTION OF DEVICES AND SERVICES.

DOI: 10.5220/0001843001050110

In Proceedings of the Fifth International Conference on Web Information Systems and Technologies (WEBIST 2009), page

ISBN: 978-989-8111-81-4

– could be overcome by using comprehensive,

semantically-rich description language.

Until the moment, there are published several

proposals for languages and systems supporting

universal service descriptions (Bandara A., 2004),

(OWL-S, 2007), (Sun, 2007). Universal frameworks

such as uMiddle (Nakazawa J., 2006) try to enable

seamless device interaction over diverse middleware

platforms. However, they fail in achieving flexible

and adequate interoperability between environments

based on different service technologies. Creation of

languages as USQL (Unified Service Query

Language) is quite prominent as this is a way to

enable service requestors to interact with

heterogeneous service registries, for the discovery of

available services (Pantazoglou, 2006).

2 REQUIREMENTS FOR

SERVICE DESCRIPTION

LANGUAGE

Service descriptions can be done in various levels of

detail. They range from simple keyword-based to

highly complex ontology-based approaches and

many different languages have been proposed in the

last years. We can distinguish four clusters of

description languages: keyword-based, template-

based, object-based and ontology-based descriptions

(Klein M., 2003). As presented in figure 1, the lower

is the semantic expressiveness the higher is the

ability of such a language for automated

comparison. While keyword-based descriptions are

far from semantic context representation, template-

based approaches - such as WSDL, SLP (Bratoev

M., 2006) and UPnP (Microsoft, 2007), and object-

based representations - like these used in CORBA

and Jini (Sun, 2007), try to structure service

descriptions into templates and objects. However,

they provide means for describing non-functional

service aspects while functional semantics have to

be still derived from textual descriptions, names and

parameters. Finally, ontology-based approaches -

such as RDFS, DAML-S (Defense, 2006) and OWL-

S (OWL-S, 2007), succeed in providing higher

expressiveness by trying to define an upper

extendible ontology for all types of services where

there are proposed general service descriptions for

both non-functional and functional service attributes

such as input/output parameters and result effect.

The definition of a powerful, protocol

independent, language for service description is

vitally important for implementation of mediated

translation. In principle, this language should be

highly expressive and able to represent every

concept in a set of real service description

languages. It must allow automatic comparison of

service descriptions without any need for additional

human analysis. It also has to be flexible and

adaptable to changes in order to reflect the

evaluation of real description formats.

Figure 1: Classification of service description languages.

A good service description language has the

following important characteristics:

• Semantically expressive - service descriptions

must be expressive and flexible, scaling to

future device types and providing support for

reasoning about services.

• Automatically comparable – the ability for

automatic comparison of service descriptions

is essential and implies well structured

descriptions

• Flexible and extensible – as the information

that should be included in the service

description changes the language must be

able to accommodate the new reality

• Editable – descriptions should be easily created

and modified by humans, possibly by using

special-purpose editors.

Our vision for such a language is that it should be

ontology-based. The particular format for service

description should be able to represent various

properties of the service such as its interface and

attributes as well as its dynamic behavior. When

examining requirements about semantic

expressiveness, we naturally come to the conclusion

that only ontology-based description languages can

provide enough expressiveness for an automatic

service trading. Describing services using ontology

is superior to using other forms of data structures

such as service templates etc. used in current

standards, because that method provides a structure

that makes it possible to reason about and derive

WEBIST 2009 - 5th International Conference on Web Information Systems and Technologies

106

knowledge from the given descriptions. These

semantically rich descriptions enable automated

machine reasoning over service and domain

descriptions, thus supporting automation of service

discovery, composition, and execution, and reducing

manual configuration and programming efforts.

Some services specifically support, or are offered

by devices (i.e. hardware components), whereby the

characteristics/capabilities of the device may play an

integral role in the description and behavior of the

services it offers. When a service involves a

hardware device (for example printing service,

scanning service) some level of detail about the

device in which it is hosted will be required for

service selection purposes.

Capabilities of the device may play an integral

role to both the description and behavior of the

services it offers. Such information relating to the

device could be included along with the service

description itself (in the service ontology), but

having separate ontologies to describe devices and

services promotes ease of use, readability and

reusability and is therefore a better design (Bandara

A., 2004). By describing devices capabilities, the

device ontology may become useful in cases of

service composition where is needed to get

knowledge about the capabilities of available

devices (such as resource capability, device

proximity to individual services, etc.) in order to

determine a broker platform.

3 THE DEVICE ONTOLOGY

There is a great variety in the way ontologies are

created, and an ongoing discussion in the ontology

community about the best practices for ontology

development. One of the greatest challenges in

constructing an ontology is the lack of formal

standards or consensual methodology. Nevertheless,

there are some processes that should be addressed.

In practical terms, developing an ontology includes

(Noy N., 2001):

• Defining classes in the ontology

• Arranging the classes (concepts) in a taxonomic

(subclass/superclass) hierarchy

• Defining slots (attributes of the classes) and

describing allowed values for these slots

• Filling in the values for slots for instances

Table 1: Device concepts.

To create the device ontology we have tried to

enumerate all the important terms for the device –

hardware, software, service, embedded device,

serial number etc. From this list we select the terms

that describe objects having independent existence

rather than terms that describe these objects. These

terms will be classes in the ontology and will

become anchors in the class hierarchy. On the other

hand, we identified all device-related properties

(table 1) that are incorporated in service description

of the three real target languages. For illustrative

purposes and without violating the general

applicability of our approach we will take three

discovery technologies (Jini, UPnP and SLP) and

carefully examine the service and device description

concepts used by each of them. These technologies

are mutually orthogonal in respect to their service

description approach (respectively, object oriented,

template-based and keyword-based) to minimize

distortion of analysis results. For each property in

the list, we must determine which class it describes.

If we cannot find semantically relevant class to add

this property (slot) to we create a new class. To

incorporate the information from table 1 in the

device ontology we created one class named

Description. This class aggregates all the

information that can be used for device discovery

and advertisement. This class has a number of

subclasses that group a set of semantically related

properties (for example, class Vendor has properties

Vendor Name and Url).

The produced device ontology is shown on figure 2.

The Service class provides the information about the

service(s) hosted on the device concerned. The

UNIVERSAL, ONTOLOGY-BASED DESCRIPTION OF DEVICES AND SERVICES

107

OWL-S could be potentially used to describe these

individual services.

Figure 2: The proposed device ontology.

4 THE SERVICE ONTOLOGY

When starting to develop ontology, it is highly

recommended to consider existing ontologies in the

same or similar domain (Noy N., 2001). Most

prominent examples of ontology-based service

languages are OWL-S (OWL-S, 2007), i.e. OWL-

based (Ontology Web Language) web service

ontology (formerly DAML-S (Defense, 2006)), and

SAWSDL (W3C, 2007):

• OWL-S markup of Web services facilitates the

automation of Web service tasks, including

automated Web service discovery, execution,

composition and interoperation (OWL-S, 2007);

• SAWSDL enables semantic annotations for Web

services not only for discovering Web services

but also for invoking them (W3C, 2007).

The most prominent ontology-oriented language

intended especially for construction of web service

descriptions is OWL-S. It is on ontology of service

concepts and provides a core set of markup language

constructs for describing the properties and

capabilities of a Web service in unambiguous,

computer interpretable form. Therefore, it provides

an upper ontology for all Web services, is extremely

generic and organizes a service description into four

conceptual areas (fig. 3):

• Service - what is the service itself?

• Process model - what does the service provide

for prospective clients?

• Profile - how is the service used?

• Grounding - how does one interact with it, by

which transport protocols?

Figure 3: Top level representation of the service ontology.

An OWL-S Profile describes a service as a function

of three basic types of information: provider

description consisting of contact information about

the service provider, functional description in terms

of the transformation produced by the service

(required inputs, generated outputs, preconditions

and expected effects) and, finally, features of the

service - category, quality rating of the service, and

an unbounded list of service parameters of any type.

The service profile is used to characterize a service

for purposes such as advertisement, discovery, and

selection. Service profiles may be published in

various kinds of registries, discovered using various

tools, and selected using various kinds of

matchmaking techniques. OWL-S 1.2 defines a

Process (i.e., a perspective on how to interact with a

service) as a subclass of ServiceModel, where

outputs and effects can depend on conditions that

hold true of the world state at the time the process is

performed. An OWL-S process may be atomic or

composite, while the control constructs may be

Sequence, Split, Choice, Condition, If-Then-Else,

Iterate, etc. (OWL-S Coalition., 2007). While both

the ServiceProfile and the ServiceModel are thought

of as abstract representations, the ServiceGrounding

deals with the concrete level of specification.

There are two main tasks in the development of

OWL-S services: (1) to define the service’s domain

ontologies (in terms of OWL classes, properties, and

instances), and (2) to create an OWL-S service

description relating to the domain ontologies and

consisting of instances of OWL-S classes such as

Service, Process, Input, and Output. Unfortunately,

OWL-S is not directly applicable for services

accessible by all existing service discovery systems.

The lack of concept for service state and events

(which are available in UPnP for example) as well as

for some service properties such as type, identifier

WEBIST 2009 - 5th International Conference on Web Information Systems and Technologies

108

Table 2: Service concepts.

and version in OWL-S ontology impose the need of

extending it in order to handle these specific

modeling situations. Table 2 below summarizes the

availability of a specific service description concept

in each of these discovery systems. It is clear that

OWL-S is not directly applicable for services

accessible by all existing service discovery systems.

The lack of concept for service state and events

(which are available in UPnP for example) as well as

for some service properties such as type, identifier

and version in OWL-S ontology impose the need of

extending it in order to handle these specific

modeling situations. Every service concept available

in one or more of the analyzed real description

languages but missing in OWL-S framework should

be associated and included to extended service

profile or extended service model producing the

following revised OWL-S service description

ontology. Since OWL-S provides a good starting

point for our work, by providing a set of concepts

for modeling basic aspects of services there are

some service concepts in the examined technologies

that are not covered by OWL-S in its pure (original)

form. That’s the reason it cannot play the role of

unified common description language that is able to

describe any service without loss of information.

Therefore, it is important to incorporate into OWL-S

information about missing concepts. That is the

reason making us to propose several enhancements

to OWL-S semantics.

The proposed extensions of the Profile and

Model schemes are not meant to provide canonical

and universal ontology for use on all service

discovery systems, but are meant merely to illustrate

technical features of this approach. Ultimately, these

enriched elements are intended to describe more

precisely the service and its functionalities. It is up

to service providers who create the service profiles

and models to decide whether to take advantage of

them and how. The unified ontology (figure 4)

together with a set of particular instances of its

classes represents a service description.

5 CONCLUSIONS AND FUTURE

WORK

In this article we proposed an ontology-oriented

perspective for device and service description in a

formal, platform-independent manner. Such a

description is essential for suppressing the

integration crisis between different service discovery

systems. The presented device ontology and OWL-S

enhancements (the new Service Profile and Service

Model subclasses) contain the all information for a

service that can be found in Jini, UPnP or SLP and is

UNIVERSAL, ONTOLOGY-BASED DESCRIPTION OF DEVICES AND SERVICES

109

Figure 4: The unified ontology for device and service description.

important for achieving a common semantic model

of a service (device). supported by the sister project

(ec fp7-sp4 capacities agr. no. 205030), we are

creating a prototype application (employing owl-s

api) that translates a service description from jini to

upnp format using our framework. the specified

extensions and the device ontology are constructed

using protégé and its open source ontology owl

editor and demonstrate the feasibility and

effectiveness of the proposed ontology. furthermore,

we plan to investigate mechanisms that enable users

to specify discovery preferences in an unobtrusive

manner.

REFERENCES

Bandara A., Payne T., 2004, An Ontological Framework

for Semantic Description of Devices (Poster), In

Proceedings of International Semantic Web

Conference (ISWC), Hiroshima, Japan.

Bratoev M., Bontchev B., 2006, Comparison of Service

Discovery Protocols, In Journal of Information

Technologies and Control, October 2006, pp 47-52.

Defense Advanced Research Projects Agency 2006

DARPA agents markup language – services (DAML-

S), http://www.daml.org

Klein M., Köning-Ries B., 2003, A Process and a Tool for

Creating Service Descriptions based on DAML-S, In

LNCS “Technologies for E-Services”, Vol. 2819,

2003, pp.143-154.

Microsoft Corp., 2007, Universal Plug and Play,

http://www.upnp.org

Nakazawa J., Yura J., Tokuda H., 2006, uMiddle: A

Universal Framework for Bridging Diverse

Middleware Platforms, White paper, Georgia Institute

of Technology, 2006.

Nakazawa J., Keith Edwards W., Ramachandran U.,

Tokuda H., 2006, A Bridging Framework for

Universal Interoperability in Pervasive Systems, In

Proc. of the 26

Int. Conf. on Distributed Computing

Systems (ICDCS), July 04-07, 2006, p.3.

Noy N., McGuinness D., 2001, Ontology Development

101: A Guide to Creating Your First Ontology,

Stanford Knowledge Systems Laboratory Technical

Report KSL-01-05 and Stanford Medical Informatics

Technical Report SMI-2001-0880

Ong S., Tang Kong., 2005, Service Description and

Composition Using Ontologies, Proc. of IASTED Int.

Conf. On Artificial Intelligence and Applications,

Innsbruck, Austria, 14-16 February, 2005.

OWL-S Coalition., 2007, OWL-S 1.2 Pre-Release,

http://www.ai.sri.com/daml/services/owl-s/1.2/

Pantazoglou M., Tsalgatidou A., Athanasopoulos G.,

Pilioura, T., 2006, A Unified Approach towards the

Discovery of Web and Peer-to-Peer Services, In Proc.

of the IEEE Int. Conf. on Web Services, ICWS 2006,

Chicago, USA, September 18-22, 2006, pp.901-902.

Sun Microsystems., 2007, Jini, http://www.jini.org

Uschold M., Gruninger M., 1996, Ontologies: Principles,

Methods and Applications, Knowledge Engineering

Review 11(2).

W3C, 2007 Semantic Annotations for WSDL and XML

Schema, W3C Working Draft, April 10, 2007,

http://www.w3.org/TR/sawsdl/

WEBIST 2009 - 5th International Conference on Web Information Systems and Technologies

110