Almaden - Austin - Beijing - Haifa - India - T. J. Watson - Tokyo - Zurich
In the last few years a paradigm shift in distributed computing could be witnessed. Service Oriented Computing has paved the way to changes in the design, architecture, delivery, and use of software applications. Services are defined as autonomous platform independent software elements which can be used to collaborate as distributed applications especially across organizational boundaries. They can be described, published, discovered, orchestrated, and programmed using standard protocols.
A promising way to realize the idea of Service Oriented Computing are Web Services. These services are designed for the implementation of business processes within a distributed computing environment and are based on a set of evolving standards.
In order to allow for PhD students to get involved into the current developments in the area of Service Oriented Computing and Web Services, the IBM PhD Student Symposium provides a forum for the presentation of PhD theses. PhD students get the opportunity to present and discuss the current status of their work in order to receive valuable feedback.
The focus of the IBM PhD Student Symposium is aligned to the co-located ICSOC conference (International Conference on Service Oriented Computing). Therefore, contributions with respect to the following topics were requested in the call for papers.
– Service Information Modeling • Service Description • Service Dependency Modeling • SLA Modeling – Service Configuration • Service Composition • Service Discovery • Service Deployment – Service Monitoring and Management • Service Performance • Service Scheduling • Service Fault Management • SLA Management – Services Framework • Service Lifecycle Management • Service Oriented Architectures • Quality of Service • Services and Workflow/Business Processes • Service Semantics – Applying Service-Oriented Architectures • Theoretical Application of Services • Services in Grid Environments • Services for Peer-to-Peer Applications • Security and Privacy Issues with respect to Services
The review of each submission has been performed by two renowned experts in the area with respect to innovativeness and contribution, relevance to call for papers, technical content, paper organization and presentation, reference to related work, and overall recommendation.
November 2005 Andreas Hanemann
Program Chair
IBM PhD Student Symposium
Organization The IBM PhD Student Symposium is held in conjunction with the 3rd International Conference on Service Oriented Computing (ICSOC 2005). It is organized as a whole day event on December 12, 2005 in the Mercure Hotel Amsterdam aan de Amstel (Amsterdam, The Netherlands). The homepage of the symposium can be found on the conference homepage (www.icsoc.org ⇒ PhD Symposium).
Program Chair: Committee Members:
IBM Research, USA Andreas Hanemann (Leibniz Supercomputing Center, Germany) Marco Aiello (University of Trento, Italy) Claudio Bartolini (HP Labs, USA) Schahram Dustdar (Technical University of Vienna, Austria) Rik Eshuis (Technical University of Eindhoven, The Netherlands) Andreas Hanemann (Leibniz Supercomputing Center, Germany) Alexander Keller (IBM Research, USA) Rania Khalaf (IBM Research, USA) Frank Leymann (University of Stuttgart, Germany) Massimo Mecella (University of Rome, Italy) Mike Papazoglou (Tilburg University, The Netherlands) Pierluigi Plebani (Politecnico di Milano, Italy) Thomas Risse (Fraunhofer IPSI, Germany)
Web Service Discovery with Implicit QoS Filtering . . . . . . . . . . . . . . . . . . . . Natallia Kokash (University of Trento, Italy) 1 7 13 19 25 31 37 43 49 55 61 Web Service Composition Quality Modelling . . . . . . . . . . . . . . . . . . . . . . . . . Ganna Frankova (University of Trento, Italy) Extending OWL for QoS-based Web Service Description and Discovery . . Kyriakos Kritikos (University of Crete, Greece) 73
Abstract. In the field of Semantic Web Services (SWS), the OWL-S and WSMF approaches provide us with ontology-based frameworks for WSDL web services. These description frameworks enable automation of high-level tasks such as discovery, invocation and composition of web services. In our work we focus on the software architectural aspects of ontology-enabled services to define an Ontologyenabled Service-Oriented Architectural style (Onto-SOA). The proposed style provides for a general integration mechanism for ontologies and SOA in a manner consistent with the SOA constraints. Onto-SOA is independent from an ontology language and a particular web service technology. Therefore, it can be easily combined with other styles to devise software architectures for existing SWS frameworks as well as novel approaches to integration of ontologies and SOA. 1
The Service-Oriented Architectural style (SOA) is gaining momentum. Gartner predicts
that SOA will become a prevailing software engineering practice in coming years1. SOA
attracts industry attention by providing for a flexible and cost effective way to re-use
functionality captured in loosely coupled, business-aligned services. WSDL [
The Semantic Web [
The significant potential of the combination of web services and the Semantic Web
has been recognized by the Semantic Web Services research area, with OWL-S [
On the more practical side, our work is motivated by the problem of applying ontologies in design of applications for the e-Science domain. In the e-Science domain a large variety of knowledge intensive computational resources exists: experiment design and execution environments, model simulation and statistical analysis services. Ontologies have a great potential to improve (re-)usability of these resources. In order to realize this potential, end-user applications have to be developed. However, at present, an application developer faces both complexity of the state-of-the-art approaches and very limited guidance available on how these approaches can be applied to development of ontology-enabled software. We address this problem by further operationalizing the proposed Onto-SOA style into a pragmatic framework ready to be applied in ontologyenabled applications.
The rest of the paper is devoted to a brief introduction to SOA (Sec. 2) followed by the analysis of relations between SOA, web services (Sec. 3) and Semantic Web Services (Sec. 3). During this analysis we outline a gap between SOA and current approaches to ontology-enabled web services. We address the gap by defining Onto-SOA (Sec. 4). Finally, we conclude with a summary and directions for future work. 2
An architectural style is “a coordinated set of constraints on architectural elements
and relationships among those elements within any architecture that conforms to that
style” [
SOA is a style that is constrained to induce a number of desirable characteristics on a compliant architecture. In the literature, SOA is often characterized as a style that supports loosely coupled, business-aligned, networked services to enable flexibility and interoperability, in a technology-independent manner.
Services in SOA represent coarsely-grained expertise from an application (business) domain. A service makes this problem-solving expertise available to a consumer. SOA achieves loose coupling between its processing components (a service consumer and a service) by requiring that connectors are simple, generic and application independent, and that descriptive messages constrained by an extensible schema are delivered through these connectors. A connector is application independent and generic if it does not introduce real dependency and minimizes artificial dependency between a service and consumer.
Since the connectors are generic, all application-specific semantics must be expressed in descriptive messages that communicate a problem description from a service consumer to a service provider. The messages specify what is to be solved but not how it should be solved. The reason for this is that a service provider possesses problemsolving expertise that is missing in the service consumer.
In order to enable “understanding” between communicating components, messages must have unified syntax and structure and must be expressed in a vocabulary shared by communicating components. The vocabulary is defined in a schema, extensibility of which is very important due to an open range of possible application domains.
In SOA, loose coupling is achieved by enforcing application independent connector elements. All application-specific semantics is to be contained in data elements (messages). A service can be uniquely characterized by a set of messages it can properly interpret. These messages are constrained by a certain schema that can be seen as an abstract definition of a service. From a service consumer point of view, a service is known via this abstract definition only. Since a service represents problem-solving expertise in an application domain, the schema effectively plays a role of a specification of this expertise. Such view on a schema in SOA allows us to make a natural transition from a schema to an ontology and treat an ontology as an integral component of SOA. 3
In the literature the term web services most often refers to WSDL/SOAP-based services. However, in our work we employ it in more general sense to refer to a derivation of SOA that employs standard Web transport protocols (HTTP, for example) and XML language to express messages. The main reason for this is that WSDL/SOAP is not the only possible way to apply SOA on the Web.
WSDL/SOAP Web Services presently are the most popular application of SOA on the
Web. WSDL provides a description framework for web services and is aimed at service
invocation primarily. SOAP [
In the very beginning SOAP supported the RPC (Remote Procedure Call) communication style only2 and later on the document-based flavor of SOAP was introduced. The document-based SOAP messages are capable of carrying any XML-based content and are not restricted to the RPC communication style only.
From the SOA perspective, RPC introduces an additional artificial dependency between a consumer and a service and, therefore, hinders loose coupling between these components. Furthermore, the RPC style messages tend to be prescriptive rather than descriptive: with the RPC-message a consumer prescribes a service how to solve a problem rather than describing what is to be solved. In RPC web services, message semantics often becomes workflow-dependent and that hinders integration of services. With SOAP 1.2 the RPC style has become optional, thus, enabling us to avoid it in WSDL/SOAP web services. However, RPC still attracts most of the attention in the research community.
Document-based WSDL/SOAP web services are fully compatible with SOA. Normally they rely on the XML Schema language to describe a structure of an XML document. However, XML schemas do not allow us to capture semantics of message ele2 http://www.xml.com/pub/a/ws/2001/04/04/soap.html ments. WSDL extension elements allow us to use different schema languages capable of capturing semantics, however, there is little guidance available on that.
WSDL/SOAP web services introduce a significant amount of conceptual and
architectural elements [
WSDL/SOAP web services require a significant infrastructure and have a high adoption barrier. On the other hand, REST web services require little infrastructure in addition to what is already provided by the Web, are fully compatible with the SOA constraints, can be easily made to employ an ontology-based schema definition languages and sufficient for many applications.
Despite the fact that REST web services are not the standard web services and less widely publicized, they are known in some cases to be preferred over WSDL/SOAP web services. For example, Amazon provides interfaces for both WSDL/SOAP and REST web services, and 85% of the usage is on the REST interface3.
Semantic Web Services The state-of-the-art approaches to Semantic Web Services [
From the SOA perspective, SWS define a relatively complex semantic connector
element responsible for establishing a connection between a service and its consumer
in a way that insures semantical compatibility of messages. The connector element
can perform complex operations, however, the service still operates on semantics-free
messages (SOAP RPC in most cases). This makes us to believe that the
state-of-theart approaches to SWS do not address direct exchange of semantically rich messages
between processing components in SOA. For example, if there is a web service designed
3 http://www.oreillynet.com/pub/wlg/3005
to operate over RDF [
We propose Onto-SOA – SOA that assumes a direct exchange of semantically rich messages between processing components. Onto-SOA is general meaning that it can be combined with any SOA-compatible architecture (WSDL/SOAP or REST) and with any ontology language. The style can be gradually refined into derivative styles that will support additional service-related activities.
We derive the Onto-SOA style from SOA by introducing an additional constraint on architectural data elements (messages): An ontology language must be used to express a schema underlying messages. This constraint explicitly addresses the semantical interoperability of messages. In Onto-SOA we assume that both a service and its consumer are aware about an ontology underlying messages.
Implementing an Onto-SOA with an Ontology Language To check whether the newly-derived Onto-SOA is compatible with an underlying SOA, we have to validate the introduced constraint against the general SOA requirement on schema extensibility.
In most of the modern applications of SOA, XML and XML Schema provide a unified syntax and vocabulary definition mechanism to messages. However, the XML Schema language addresses structural aspects only, leaving semantics of a defined vocabulary implicit. This leaves a problem of semantical interoperability among processing components unsolved but, on the other hand, does not restrict the flexibility in defining domain-specific vocabularies.
An ontology can provide a domain vocabulary, semantics of which is precisely
defined in terms of ontological primitives of the underlying language. The ontology
languages such as RDFS [
This allows us to conclude that in order to achieve the level of flexibility required by SOA, semantics of an ontology language must be extensible. However, at present, such languages are not available and we cannot foresee whether they will appear in future. If an extensible ontology language is unavailable then the least restrictive one (such as RDFS, for example) might provide a fair approximation.
In order to operationalize the introduced Onto-SOA, we have devised MoRe – a
derivation of Onto-SOA (and a corresponding implementation) that combines the RDF/S
languages with the elements of REST web services. MoRe has been applied to provide
a number of services for the domain of units of measure captured in the UnitDim4
ontology. The services are employed within the Unit Converter tool [
http://www.atoapps.nl/foodinformatics. Sec. News (2004)
We propose Onto-SOA – an Ontology-enabled Service-Oriented Architectural style that addresses a direct exchange of semantically-rich (ontology-based) messages between a service and its consumer. Onto-SOA is independent from a particular web service technology (WSDL/SOAP, for example) and, therefore, can be employed within any SOAcompatible architecture. Onto-SOA makes no assumptions about an ontology language employed as a schema definition mechanism. This enables us to combine Onto-SOA with any ontology languages.
In our future work we intend to further analyze the applicability of the modern ontology languages such as RDF/S and OWL within the proposed Onto-SOA style. Further applications the MoRe framework – an operational RDFS/HTTP derivation of Onto-SOA – within the e-Science domain will allow us to validate the main ideas behind Onto-SOA and to provide additional guidance to software and ontology developers in designing an ontology- and service-enabled architectures.
Engineering Authorization Services for the Service Oriented Architecture
Sarath Indrakanti Information and Networked Systems Security Research,
Department of Computing, Macquarie University,
Sydney, NSW, 2109, Australia
sindraka@ics.mq.edu.au Abstract. The service-oriented architecture (SOA) can be used to build new solutions leveraging services, to cleave together existing applications or to cleave apart existing applications. The SOA provides many benefits such as cost saving to organizations by increasing the speed of implementation of any application(s) required and reducing the expenditure on integration technologies. However, security is one of the main roadblocks for enterprises to delay development and deployment of their services. Although there are standards for providing confidentiality, integrity and message authentication for services, there is not yet a standard specification for authorization services for the SOA. We address this important gap in the area of security for the SOA. In particular, we will propose an authorization policy language as well as an authorization framework for the SOA. 1
Background The SOA is an architectural style and its aim is to achieve loose coupling among interacting distributed software systems. The SOA can be defined as a way of designing and implementing enterprise applications that deals with the intercommunication of loosely coupled, coarse grained, reusable artifacts (services). The SOA is made up of independent services interconnected via messaging. Platform independent service interfaces are defined to invoke these services. The SOA consists of service providers and service consumers. Service providers define what the service looks like and how to invoke it through an implementation independent service interface. Service consumers use this interface to invoke the service. The SOA also provides discovery mechanism to act as an intermediary. Service providers publish their service interface using the discovery mechanism for consumers to find and invoke the service.
There are three broad categories of service use. 1) To build new solutions with services. Services enable enterprises to have independent pieces within applications that can be developed and maintained independently and also to scaleout. 2) To cleave together existing applications. Services enable connectivity Business-to-Business (B2B) and Enterprise Application Integration (EAI). Services also let us formulate business processes to enable workflows across heterogeneous environments in turn letting us tap in to the IT infrastructure, and 3) To cleave apart existing applications. Use cases of the SOA include Supply Chain Management, Customer Relationship Management (CRM), Enterprise Application Integration (EAI), Enterprise Resource Planning (ERP), and Portal Web Sites amongst others. The SOA enables cost saving by increasing the speed of implementation of any application(s) required and reducing the expenditure on integration technologies. The SOA provides enterprise agility, to quickly respond to rapidly changing market needs and business requirements by quickly building new applications and quickly updating old applications. 2
Securing the SOA In general, security for the SOA is a broad and complex area covering a range of technologies. At present, there are several efforts underway that are striving to provide security services such as authentication between participating entities, confidentiality and integrity of communications. A variety of existing technologies can contribute to this area such as TLS/SSL and IPSec. There are also related security functionalities such as XML Signature and XML Encryption and their natural extensions to integrate these security features into technologies such as SOAP and WSDL.
The WS-Security specification describes enhancements to SOAP messaging
to provide message integrity, confidentiality and authentication. There is also
work on XKMS defining interfaces to key management and trust services based
on SOAP and WSDL. However, while there is a large amount of work on general
access control and more recently on distributed systems authorization, research
in the area of authorization for Web services is still at an early stage. There is
not yet a specification or a standard for Web services authorization. There are
attempts by different research groups to define authorization frameworks [
Authorization Policy Language for the SOA [
In particular, we will discuss the following in the dissertation: (a) Survey of
authorization policy languages for distributed systems, (b) Analysis of the
authorization policy language used by .NET MyServices [
Survey and Analysis of Authorization Frameworks[
We will discuss the following in this section: (a) Survey and analysis of authorization frameworks built for stand-alone systems, (b) Survey and analysis of authorization frameworks built for distributed systems, and (c) Survey and analysis of authorization frameworks built for the SOA. 5
Authorization Framework Design Requirements [
Authorization requirements differ for the Web services and business processes
layers of the SOA. Authorization services for the Web services layer have special
design requirements as Web services present a complex layered system. For
instance, a service may be a front-end to an enterprise system and the enterprise
system may access information stored in databases and files. Web services may
be used by enterprises to expose the functionality of legacy applications to users
in a heterogeneous environment. Or new business applications could be
written to leverage benefits offered by Web services. This means an authorization
architecture for Web services must support multiple models of access control.
This enables legacy applications to use the access control models they have
already been using as well as new Web services applications to use new models of
access control or other well-known models of access control such as role-based
access control (RBAC). An authorization architecture for the business process
layer of the SOA must provide orchestration services to coordinate the
authorization decisions from individual partners authorization policy evaluators. Each
partner must be allowed to control its own authorization policies and also not
require disclosing them to the entire workflow or to the workflow engine. Even
in cases where the binding to actual end-points of partner services happens
dynamically at runtime, the authorization architecture must be able to orchestrate
the partners authorization policy evaluators and arrive at an authorization
decision. Currently existing authorization frameworks are either designed for the
Web services layer [
Authorization Framework for the SOA We will propose a unified authorization framework for the SOA. The framework will provide two separate architectures (indicated by the light grey colored boxes in Figure 1) that extend the security layers of Web services (WSAA) and business processes (BPAA). The BPAA relies on the features provided by the WSAA. In particular, we will discuss the following: 6.1
Web Services Authorization Architecture (WSAA) [
Business Processes Authorization Architecture (BPAA) (a) Design of the authorization administration and runtime APIs. The administration API is useful to define a business process and mange its authorization related information. The runtime API is useful to invoke the necessary authorization components and receive an authorization decision, (b) Implementation of the authorization architecture APIs within the .NET framework, and (c) Benefits of the proposed architecture.
Extensions to Other Layers of the SOA We will also describe our extensions to the Web services description and messaging layers to support the authorization frameworks for Web services and business processes (indicated by the dark grey colored boxes in Figure 1). Extensions to the description layer are required to make prospective clients aware of what authorization privileges are required to invoke a Web service or a business process. Extensions to the messaging layer are required to carry authorization related privileges, for instance, in the form of credentials. 7
Demonstration of Authorization Services in a Health Care Application We will demonstrate the proposed authorization services the policy language and the authorization framework for the SOA using a case study in the health care domain. In particular, we will discuss the following:
(a) Demonstration of the proposed policy language and policy engine features, (b) Demonstration of the authorization services provided by WSAA and BPAA, and (c) Demonstration of the proposed extensions to the service description and messaging layers.
Conclusion The service-oriented architecture (SOA) can be used to build new solutions leveraging services, to cleave together existing applications or to cleave apart existing applications. The SOA provides many benefits such as cost saving to organizations by increasing the speed of implementation of any application(s) required and reducing the expenditure on integration technologies. However, security is one of the main roadblocks for enterprises to delay development and deployment of their services. Although there are standards for providing confidentiality, integrity and message authentication for services, there is not yet a standard specification for authorization. We will address this gap in the area of security for the SOA. In the dissertation, we will first address the area of authorization policies and describe our proposal of an XML-based authorization policy language and its evaluation engine for the SOA. Then we will discuss the authorization framework requirements for the SOA. They differ for the Web services and business processes layers comprising the SOA. We highlighted the major differences in this paper. In particular, we will describe two separate authorization architectures designed using the requirements we laid out after carrying a comprehensive survey of authorization models built for stand-alone and distributed systems as well as for the SOA. Finally, we will demonstrate the authorization policy language and the authorization framework for the SOA using a case study in the health care domain.
Web Services Software Architecture
Syahrul Fahmy School of Informatics, The University of Manchester, PO Box 88, Manchester M60 1QD, United Kingdom
S.Abdul-wahab@postgrad.manchester.ac.uk Abstract. Web services have received widespread attention and acceptance in the software engineering community. Automatic composition of services, to meet user’s requirements, is a powerful mechanism to ensure satisfaction of varied needs and thus enable the vision of Web services. Although Web services are equipped with the fundamental concepts and supporting technologies, the architectural style of the composite service composed from Web services is fixed by the prescriptions of the Service Oriented Architecture, and does not provide the level of flexibility stipulated by alternative service-based approaches such as the Service-Based Software vision. To enable the composition of Web services using alternative architectural styles, the author proposes that software architecture be included as a parameter in the composition process. This is enabled by the use of the Alfa framework to model the desired architectural style, and the use of this information to support service discovery. 1 Introduction Web services are configurable software services that use open standards and protocols to connect and integrate distributed components for creating and managing computer applications. Web services have been becoming a major software trend because they provide means for integrating data sources and data transfer between applications. Web services share business logic, data and processes through a programmatic interface across a network. In order for web services to be operational, interoperability of applications over intra/internet is necessary. This interoperability requires a technology that is independent of hardware, operating system, and software specific technologies. As such, specific standards are used to address these interoperability issues that consist of XML to define structured data, SOAP as a messaging protocol, WSDL to define interfaces, and UDDI as a registry to publish available services.
There are two approaches an organization can adopt to take advantage of Web
services namely the Service Oriented Architecture (SOA) and the Service-Based
Software (SBS) approaches. In the SOA approach, a fixed architecture is presumed for
service composition. Little flexibility, if any, is available to the service provider in
determining how the service should be composed. The author argues that the service
provider should be given more flexibility and ‘authority’ in service composition. In
the SBS approach, the service provider is given this flexibility by means of a flexible
– Protocol algebra and protocol management operators. A distinctive feature of our
approach lies in the definition of operators to query, analyze, and transform
protocols. Such operators are the key to carry out most of the features described above.
– Protocols adapters and code generation. Theses two issues are part of the
ServiceMozaic platform but won’t be further detailed in this paper. Briefly, adapters [
State of the art
Tools supporting web services development today are mainly concerned with
interoperability issues at the lower levels of the web services stack, like the mappings from
WSDL descriptions to Java/C# source code and vice-versa, making two SOAP-based
systems talk to each other. Similarly, the existing standards in the higher-level services
descriptions such as BPEL4WS, WSCI or WSCL proved to be more concerned with
implementation aspects than enabling the kind of formal analysis that we envision.
Indeed, the importance of formal analysis of web services protocols in terms of automated
support for services interoperability at the business protocol level has been discussed
in recent papers: [
Our approach, based on a protocol algebra and protocol operators is novel in the
field. Specifically, it should be more fine-grained than the approaches mentioned above
when doing compatibility and replace-ability analysis. We believe that to some extent,
the development of a protocol algebra for formal services descriptions analysis
implemented inside a larger CASE tool, based on the identification of abstractions needed by
practitioners [
Web services protocols modeling and analysis
A model extended with temporal constraints
Our model is based on the web services business protocol proposed in [
We have extended the model (called timed web service business protocol) to cater
for temporal abstractions in [
Temporal compatibility and replace-ability analysis
Reactive systems have been widely studied, notably to the benefit of software and
hardware verification. For analysis purposes, two formal frameworks are of interest. The
first one is the temporal logics theory [
Future work includes focusing on the third kind of protocols-based analysis: consistency analysis. We will also need to take into account multi-protocols choreography 2 See http://www.eclipse.org/ and http://www.eclipse.org/gef/. analysis (the current work is focused on the analysis between a service provider and a requester). Finally, other useful abstractions such as transactional properties will provide another promising area of investigation.
Web Service Discovery with Implicit QoS Filtering
Natallia Kokash DIT - University of Trento, Via Sommarive, 14, 38050 Trento, Italy
email: natallia.kokash@dit.unitn.it
Abstract. Web Service (WS) discovery is a critical problem
hindering web service technology proliferation. The current solution, based on
catalog-style browsing, provides no control over the quality of registered
services. Application of matching techniques for WS retrieval is still
under investigation. The objective of this work is the design of a framework
to improve WS discovery. Our approach is based on application of
distributed recommendation system to provide Quality of Service (QoS)
information and on testing of retrieval methods on service specifications.
1
WS paradigm is a promising model of software technology, based on loosely
coupled, distributed and independent services operating via the web
infrastructure. To overcome platform and language dependence, services are described
using Web Service Description Language (WSDL). Standardized XML-based
interfaces help performing service reuse. Service descriptions are cataloged in
Universal Discovery, Description and Integration (UDDI) registries. Although
there exists a stack of standards to regulate the communication of processes and
automated tools to convert legacy applications into web services, WS technology
is still not widely used. One of the reasons is the lack of means to support WS
discovery, i.e., the identification of existing WSs that can be used by new web
applications. This problem is rather extensive and admits various interpretations
[
The rest of the paper is organized as follows. In Section 2, we discuss different
aspects of the problem and present the existing work relevant for our research.
Section 3 describes our approach. In Section 4, we draw some conclusions and
outline future work.
2
We can judge how well a web service satisfies a client’s goal using various types of
matching. Signature matching considers only function types without regarding
their behavior. Specification matching is a way to compare two software
components, based on descriptions of their behaviors. Components can be compared
with various degrees of accuracy (exact and relaxed matching). However, this
approach requires formal pre/postcondition specifications. Hausmann et al. [
From the WS discovery perspective, the major criterion of WS retrieval system evaluation is the relevance of found services with respect to the end user. The most popular quality evaluation measures of Information Retrieval (IR) engines are precision (a measure of the usefulness) and recall (a measure of the completeness). They have a fixed range and are easy to compare across different queries and engines, but they do not account for the quality of ranking. For a machine learning algorithm we need an effective single number measure. An average precision that combines precision, relevance ranking, and overall recall is an ideal measure for our task where good ranking is extremely relevant due to invocation cost of the services.
Assuring the WS quality is an especially critical task since we do not know
service exact specifications, developing and testing methodologies. An invocation
of a troublesome service can affect all the system. A potential client should be
aware of the service behavior before he/she/it starts to exploit it. Among the
basic QoS factors are service performance (throughput, response time, latency,
transaction time), availability, accessibility, reliability, scalability, exception
handling, execution cost, reputation, regulatory, accuracy, integrity,
interoperability, security (authentication, authorization, confidentiality, traceability, data
encryption, non-repudiation), privacy, network-based factors (network delay, delay
variation, packet loss), etc. [
An important issue is the way in which a client asks for a service. Currently
used by UDDI API keyword-based queries are easy to express but obviously
not sufficient. Using WSDL specifications as a query appears more promising.
In the case when a client wants to substitute one web service with another
without affecting the observable behavior of the system, he/she does not need
to write a request. But in general, WSDL specification is redundant and does
not contain QoS requirements. Ran [
In IR approaches to WS discovery a query consists of keywords, which are
matched against the stored descriptions in the UDDI. The Boolean model
suffered from either lots or very few returned results. Traditional models like
Vectorspace model seem to be quite effective. Latent Semantic Indexing (LSI), the
prevailing method for small document collections, was applied on the UDDI to
capture the semantic associations between services [
Retrieval is successful when the information need is satisfied. Therefore, we
need to be provided with the user’s judgement of the relevance of the found
data. Maximillien et al. [
A Framework for WS Discovery
In Fig. 1 a basic architecture of our proposed framework is presented. Each client or group of local clients with similar preferences has a dedicated agent which task is to process all activities pertaining to web services (communication with registries, bindings, requests, responses). The personal agent accepts a request from its client (human or application), redirects the description of the goal to the matching agent which manages the registry. The matching agent searches for the services that can conceivably satisfy the client’s goal. It keeps history of requests, and provides the requester with the information about the agents that searched for the similar services before. Given a list of previous clients, that is the last k agents to which the matching agent recommended the service, the personal agent can ask them for recommendations. Then two scenarios are possible: a first one, in which the agents that invoked the services provide the requester with their own rankings, and an alternative one, in which they render to it parts of their own history leaving to the requester the task of processing it. The latter approach tries to overcome the problem of specific client’s preferences that cannot be provided by external QoS evaluation systems. The personal agent can ask first the agents that gave useful statistics before, and only if they do not have necessary information, to apply more extensive interrogation. An open issue is preventing malfeasant agents from affecting the service reputation. If web services use mediators that collect statistics about all their invocations, the personal agent may have more reliable data by comparing parameters provided by agents and the service mediator. Another advantage is that centralized data is more convenient for evaluation of such QoS factors as availability, reliability, capacity, robustness.
The matching agent provides the personal agent with the matching score of its request and returned web services. If this score is low, the relevance of the services is dubious and alternative approaches are needed to check if they correspond with the client’s request. Recording of invocation inputs and outputs can help to test the service before usage by new clients. A client sends testing samples to his/her/its agent, and the agent checks if they are among the data it has. If it is so we may strongly believe that the service fits the client’s goal and will work properly. But this approach conflicts with security and privacy issues, and may be time and space consuming.
The personal agent ranks the web services by combining the scores of matching, recommendations and tests (if they were provided). Before service invocation it asks the client for confirmation. Finally, the best-fit web service is invoked, QoS parameters are measured, and the results are saved. Then, the agent can ask its user if it should always invoke this service for the same query. Such an explicit user feedback can be used to reconstruct the trust policy to other agents or rating algorithm but it is not necessary. If a user keeps invoking a web service, he/she is likely to be satisfied with the service provided. Therefore, repeated usage should have a high weight for QoS evaluation. 4
Conclusions and Future Work In this paper we have presented a new WS discovery approach. Some subproblems, namely service matching methods, discovery system evaluation and request formats have been discussed. We have proposed a framework for WS selection based on QoS properties collected implicitly by a distributed agent-based system.
We intend to apply the recommendation system presented in [
In [
Our architecture is similar to [
Ganna Frankova DIT - University of Trento, Via Sommarive, 14, 38050 Povo-Trento, Italy email: ganna.frankova@dit.unitn.it Abstract. The critical issue of web services success is the ability to compose web services in order to build complex added-value services. In web service compositions both functional and non functional properties, i.e. quality of service, should be taken into account. The quality of service in web service compositions plays a vital role and has opened a wide spectrum of challenges. We survey and analyze various approaches for modelling web service quality composition. Also we overview some ideas of service selection according to the surveyed models. From the analysis it turns out that there are many open issues in the web service quality composition area. 1 One of the most thought provoking issues in web services (WS) is that of automatic composing services in order to build complex added-value services. In particular, the quality of service (QoS) compositions plays a vital role and has opened a wide spectrum of challenges.
Nowadays the attention is on quality driven web service discovery [
In this paper, we survey and analyze various approaches for modelling web service quality composition. Some ideas of service selection according to the models are presented. From the analysis it turns out that there are many open issues in the quality of web service composition area.
The reminder of the paper is organized as follows. Section 2 discusses related work. Section 3 is devoted to the comparison of the various approaches to WS quality composition modelling with respect to the proposed requirements. Some open issues and concluding remarks are presented in Section 4. 2
Approaches for WS Quality Composition Modelling
Web service QoS issues are gaining attention and have been addressed in a
number of recent works. Some approaches are based on the extension of the Web
Service Description Language (WSDL) [
In [
In [
The feasibility of using constraint programming to improve the automation
of web services procurement is shown in [
An overview of several of the proposals for web service quality composition modelling follows. For all these approaches we present a realistic car loan example that involves getting a loan and buying a car. We assume that standardized web services interfaces for getting a loan and buying a car already exist. Some ideas of service selection according to the models are also presented.
Continuous-time Markov chains solution. The approach of estimation
execution time and cost of a workflow based on continuous-time Markov chains
is proposed in [
Quality Vector solution. Description of elementary service quality as a
quality vector is proposed in [
Applying the approach to the car loan example, a quality vector can be defined as q(s)=(qprice(s), qduration(s)), where qprice(s) is price of the service execution; qduration(s) is the expected delay in seconds between the request to the service s sending and the result receiving. The quality vector of a composite service’s execution plan p is the following: Q(p)=(Qprice(p), Qduration(p)), where Qprice(p) = PiN=1 qprice(si); Qduration(p)=CPA1(qduration(s1), ..., qduration(sN )); N is a number of states in the execution path.
Selection of optimal component services of a composite service can be done with a global planning approach. Furthermore service selection can be formulated as an optimization problem which is solved using linear programming.
Agent-oriented solution. The use of the agent-oriented methodology
Tropos to model a wide spectrum of quality of web services properties is proposed
in [
Ontology-based solution. E.M. Maximilien and M.P. Singh proposed an
ontology-based framework for dynamic web services selection [
Figure 1 shows the application of the approach to the running example. Both
for TakeLoan and for BuyCar services service and QoS ontology are presented.
1 Critical Path Algorithm [
Dynamic service selection calls for an agent-based solution. Using QoS ontology, service agents match advertised quality levels for the consumers with specific QoS preferences. 3
Discussion Now we can compare the various approaches for WS quality composition modelling with respect to the requirements derived from our exploration of the approaches we surveyed. The table in Figure 2 summarizes the results. The approaches are presented in the columns while the rows show the requirements.
Objective QoS Subjective QoS Run-time support QoS assignment
Requirements considering level
Markov Chains Quality Vector Agent-oriented Ontology-based √ √ √ √ √ Low Average
High
Fig. 2. Comparing approaches for WS quality composition modelling
Objective and Subjective QoS. QoS can be objective, such as reliability,
availability, and request-to-response time, or subjective, i.e., focusing on user
experience [
Run-time support. According to the fact that some QoS metrics such as response time can change at run-time and a single value is not appropriate, approaches for WS quality composition modelling should have a run-time support infrastructure. Markov chains- and agents-based approaches provide run-time support.
QoS assignment to composite service. When quality constraints and preferences are assigned to composite services rather than to component services it is easy to reuse the composed service, i.e., use it in another quality driven composition. The table in Figure 2 shows that quality vector-based and agentoriented solutions allow to assign QoS to composite service.
Quality Requirements. QoS Requirements for Web Services [
Summarizing, an analysis of the table shows that current QoS models of WS composition are far from ideal, and a number of open challenges exists. In this paper we have presented a survey and analysis of various approaches for modelling web service quality composition. Web service quality composition approaches range from industry standards to more abstract models. An ideal approach should satisfy the identified requirements, but so far there is not an approach fulfilling all requirements.
The survey has highlighted that there is a lot of space for further investigation and innovative research. In particular, web services are widely-used in e-commerce, an area required achieving high level of security. But there is no solution to model security properties of composed services. Further investigation should be devoted to solving security quality integration, as well as run-time support issues. Mathematical modelling approaches, and graph-based solutions are promising fields to look for solutions to this problem, as well as optimization techniques. The author will focus on solving the web service security composition problem by applying optimization techniques.
Acknowledgments The author thanks Marco Aiello and Fabio Massacci for discussions and assistance in the presented work.
Extending OWL for QoS-based Web Service Description and Discovery
Kyriakos Kritikos University of Crete, Heraklion, Greece,
kritikos@csd.uoc.gr 1
Main Theme of Thesis Web Services (WSs) are modular, self-describing, loosely-coupled, platform and programming language-agnostic software applications that can be advertised, located and used across the Internet. They are viewed as one of the promising technologies that could help business entities to automate their operations on the web on a large scale by automatic discovery and consumption of services. Based on the above reasons, the WS paradigm is being adopted by many companies and individuals and many WSs are being deployed and running.
However, as all of these WSs are advertised in a UDDI-based repository, an unavoidable fact as UDDI is a de-facto standard, the problem of discovering them based on a requester’s functional needs becomes crucial. UDDI uses a syntaxbased approach for WS description leading to purely syntactic discovery efforts returning imprecise and inaccurate results. OWL-S [OWL-S Coalition 2003] and similar joint Semantic Web and WS efforts solve the problem of syntactic WS description by using ontologies for describing WSs. Ontologies provide meaning to concepts and relationships between them, leading to semantic WS Discovery algorithms, which provide more precise and accurate results.
But even if all the advertised WSs satisfying a requester’s functional needs are returned, many results may be produced. So a non-functional concept is needed that will differentiate between the functionally equivalent WS advertisements. This concept is quality of service (QoS). QoS is closely related with the performance of a WS as well as with other features of a WS that bear on its ability to satisfy stated or implied needs. Therefore it has a substantial impact on users’ expectations from a service. Thus WS descriptions must be enhanced with QoS descriptions. Additionally, WS discovery algorithms should perform QoS-based matchmaking and selection in order to produce fewer ranked results.
Unfortunately, all the current research efforts fail in correctly describing QoS for WSs. Semantics seems to be missing from the QoS description of a WS leading to purely syntactic QoS-based WS matchmaking and selection algorithms. But even if semantics is introduced, QoS description is not rich enough and not quite extensible. So the main issue of this PhD thesis is the rich, extensible, and semantic description of QoS for WSs. Additionally, new QoS-based WS matchmaking and selection algorithms must be devised or the current best should be extended in order to take advantage of this enhanced semantic QoS description.
Motivation for Research After reviewing several definitions for QoS for WSs, we consider QoS for a WS as ”a set of non-functional characteristics/attributes that may impact the quality of the service offered by the WS ”. If a WS is advertised to have certain values (or range of values) in these QoS attributes, then it is said that this WS conforms to a certain QoS level. In this section we explain the reasons for incorporating QoS in WS description.
According to [Cardoso et. al. 2004], several researchers have identified Web Processes (WPs) as the computing model that enables a standard method of building Web-services applications and processes to connect and exchange information over the Web. For organizations, the ability to characterize WPs based on QoS has four distinct advantages. First, it allows organizations to translate their vision into their business processes more efficiently, since WPs can be designed according to QoS metrics. For e-commerce processes it is important to know the QoS an application will exhibit before making the service available to customers. Second, it allows for the selection and execution of WPs based on their QoS, to better fulfil customer expectations and requirements. Third, it makes possible the monitoring of WPs based on QoS to assure compliance both with initial QoS requirements and targeted objectives. QoS monitoring allows adaptation strategies to be triggered when undesired metrics are identified or when threshold values are reached. Fourth, it allows for the evaluation of alternative strategies when adaptation becomes necessary. It is essential that the services rendered follow customer specifications to meet their expectations and ensure satisfaction. Customer expectations and satisfaction can be translated into the quality of service rendered. Organizations have realized that quality of service management is an important factor in their operations.
As WPs are composed or single WSs, all the above advantages of QoS management of WPs also apply to WSs. So WSs can be designed and implemented according to QoS metrics (properties). They can also be discovered and selected based on their QoS capabilities. In addition, they can be monitored in order to reassure the promised QoS levels to the customers. Moreover, monitoring of QoS for WSs can trigger adaptation strategies when undesired metrics are identified, threshold values are reached, network or software or hardware errors happen. Now, we will closely examine the advantages of QoS description (management) in other non-basic activities/functions of the Service Oriented Architecture.
After the process of WS Selection, the requester chooses the best WS from an ordered WS advertisement list. However, even if WS clients find the appropriate WS, they are not confident that the WS’s described QoS levels will actually be delivered during WS execution. For this reason, the WS client and provider enter a multi-step negotiation phase, where they try to agree on a trusted third-party entity monitoring QoS levels delivered, on the penalties that will be imposed when one of the two main parties does not keep up with its promises, and on the validity period of the promises. The result of this negotiation phase is a contract or a Service Level Agreement (SLA) document that will give confidence and trust to the entities providing and consuming the service and will lead and guide the process of WS Execution. If agreement is not met, the negotiation is stopped and the WS client contacts the next WS from the returned list of the WS Selection phase.
When composing a WS, component services are associated to the individual tasks of the composite WS and are invoked during each execution of the WS. However, the number of services providing a given functionality may be large and constantly changing and some of these services will not always be available due to network problems, software evolution and repair, and hardware problems. One solution to this problem is given at design time by QoS-based WS discovery. Another solution is the runtime selection of component services, during the execution of a composite WS, based on quality criteria (i.e. constraints and preferences) and following a local [Benatallah et. al. 2002] or a global [Zeng et. al. 2003] selection strategy. In the latter case, service selection is formulated as an optimization problem and linear programming is used to compute optimal execution plans for composite services.
Review of Related Work In this section, the current research approaches for QoS-based WS Description and Discovery are described and their deficiencies are analyzed.
The Web Service Description Language (WSDL) and UDDI WS standards are syntactical approaches that do not express the QoS aspect/part of WS Description. While OWL-S is a standard semantic approach for WS Description, it does not describe QoS offers or demands as it only contains an attribute used for rating a WS.
[Tosic et. al. 2002] argue that for the specification of constraints for QoS metrics/attributes, five ontologies must be developed from which the most important (the top one) is the metrics ontology. They describe the structure and involved elements in four out of the five ontologies but they did not develop any ontology. In addition, the requirements specified are incomplete as each from the four aspects of QoS description needs further analysis.
In [Shuping Ran 2003], an extension to UDDI is proposed that represents description of QoS information about a particular WS. However, there is no actual description of the contents of this extension apart from its structure. Moreover, it relies on the UDDI (model), so it can be used only for syntactic matchmaking of offers and demands.
In [Maximilien and Singh 2002], an architecture and a conceptual model of WS reputation (QoS) (which encloses a QoS attributes model) are presented. However, the reputation of a WS is calculated and not it’s QoS. In addition, not only concepts like QoS constraints and QoS offers and demands are not modeled but also the QoS metrics model is not rich enough.
Work described in [Tosic et. al. 2003], which presents the Web Service Offerings Language (WSOL), proposes that a WS must offer different classes of service in order to satisfy a greater amount and type of customers and in order to deal successfully with situations where there is a variation in QoS due to network problems or mobility reasons. This work comes with the following shortcomings: (a) no separation and integration of constraint dimensions; (b) no specification of a QoS demand; (c) the metrics ontologies are not yet developed.
The research effort described in [Tian et. al. 2003] analyzes what must be enclosed into the QoS information for a WS request or advertisement with the help of a QoS ontology. However, not only there is not a complete and accurate description of QoS constraints but also metrics ontologies are not developed but just referenced.
In [Zhou et. al. 2004], DAML-S WS description language is extended to include a QoS specification ontology. In addition, a novel QoS matchmaking algorithm is proposed, which is based on the concept of QoS profile compatibility. The deficiencies of this research effort are the following: (a) The metrics model is not rich enough; (b) QoS metrics have the set IN+ as their range; (c) DL reasoners are slow and do not support the most complex mathematical expressions.
The research effort described in [Mart´ın-D´ıaz et. al. 2003] uses a symmetric QoS model expressing mathematical constraints for QoS metrics. However, semantics is missing leading to syntactic matchmaking and selection algorithms. Before matchmaking, a QoS specification is transformed to a Constraint Satisfaction Problem (CSP) [Hentenryck and Saraswat 1996] which is checked for consistency (if it has any solution). Matchmaking is performed according to the concept of conformance (if every solution of offer is a solution of demand). Concerning WS Selection, the (QoS) score of a WS advertisement is expressed as a Constraint Satisfaction Optimization Problem (CSOP) [Hentenryck and Saraswat 1996] where from all solutions to the CSP of an offer we try to find the one that minimizes the weighted sum of the weight of each metric multiplied with its utility assessment value. Unfortunately, CS(O)Ps can have non-polynomial solutions when there are non-linear expression at QoS constraints. 4
Future Work of the Thesis Based on the previously analyzed research work, we propose the following steps that must be taken to address the issues of the thesis; the fulfilment of which will lead to correct, efficient and accurate QoS-based WS Description and Discovery processes. Some of them have already been performed while the other will be dealt with in the future. These steps are:
Requirements for QoS-based WS Description: This step has already been taken. We have come up with the following requirements: – Extensible and formal semantic QoS model – Standards compliance – Syntactical separation of QoS-based and functional parts of service specification – Both requester and provider QoS specification – Refinement of QoS specifications (extensibility, reusability) – Fine-grained QoS specification (QoS specs for the whole WS and its parts) – Extensible and formal QoS attribute/metric model which must at least specify: (a) The value set of the attribute; (b) The domain of knowledge of the attribute; (c) The relationship of the attribute with other attributes; (d) The association of the attribute with a unit, a measured property and a measurement function; (e) A functional description of how a QoS attribute of a complex WS can be derived from the corresponding QoS attribute of the individual WSs that constitute the complex one. – Classes of service (an advertisement should present many offers)
QoS-based WS Description : Based on the above requirements, a QoSbased WS ontology has been developed with the name OWL-Q. This ontology extends OWL-S (standard) WS Description ontology and is carefully separated into many facets, each capturing one aspect of QoS WS (metric) description except from the basic one that associates OWL-S elements with OWL-Q elements. This ontology satisfies the above requirements and enables the semantic matchmaking of QoS advertisements and offers.
Semantic QoS metric matchmaking: Based on the OWL-Q ontology, we have developed an algorithm that semantically matches QoS metrics of demands and offers. This algorithm can be used as a building block for the QoS-based WS matchmaking and selection algorithms. For matchmaking simple metrics, this algorithm compares the type and measurement directive of the metrics, while for composite metrics it also compares the metrics’ measurement functions.
Develop a new or extend an old QoS-based WS matchmaking algorithm: We have extended the matchmaking algorithm proposed in [Mart´ın-D´ıaz et. al. 2003] in order to incorporate the semantic QoS metric matchmaking algorithm. To be more specific, when transforming QoS offers and demand to CSPs, we are careful to assign same metrics to same CSP variables and to perform unit transformation of the same metrics.
Develop a new or extend an old QoS-based WS selection algorithm: We have extended the selection algorithm proposed in [Mart´ın-D´ıaz et. al. 2003] in order to incorporate the semantic QoS metric matchmaking algorithm. To be more specific, when transforming QoS offers to CSOPs, we are careful to assign same metrics to same CSP variables and to perform unit transformation of the same metrics. We have also changed the scoring function: now this function is the partial sum of the minimum and maximum assessment sums. That is we solve two CSOPs for the same offer and then we perform a partial sum of their results.
Implement these matchmaking and selection algorithms: This is a step under development. The implementation uses an OWL inference engine for the semantic QoS metric matching and the ECLiPSe [Novello and Schimpf 1999] engine for solving CS(O)Ps.
Formal evaluation of the above algorithms: The above three algorithms should be formally evaluated in order to prove that they are efficient, quick, accurate and precise.
Extend the ontology and the algorithms: After performing all the above steps, OWL-Q ontology should be extended to include other non-functional descriptions of WSs (mainly contextual ones) and its design must be finalized. In addition, the QoS-based matchmaking algorithm must be extended in order to distinguish between soft and hard non-functional constraints.
Tools: GUIs and other utilities should be constructed that will help the user in describing and discovering WSs.
References