Home
QoS-based web service description and discovery
Results - Details
|
||||
| Identifier | 000339015 | |||
| Title | QoS-based web service description and discovery | |||
| Author | Κρητικός, Κυριάκος Εμμανουήλ | |||
| Thesis advisor | Πλεξουσάκης, Δημήτρης | |||
| Abstract |
Unfortunately, these two languages rely on syntactic descriptions of WS functional and non-functional attributes. This situation creates two problems. Firstly, syntactic discovery efforts return inaccurate results. Secondly, no means are provided in order to select among multiple WSs that perform the same function. The first problem has been solved by combining /Semantic Web/ (SW) and WS technologies. The second problem can be solved by taking into account a big subset of all non-functional properties of WSs, which is called /Quality of Service/ (QoS). QoS is closely related with the performance of a WS so it has a substantial impact on user's expectations from a WS and can be used as a discriminating factor between functionally equivalent WS advertisements. Thus WS descriptions must be enhanced with QoS descriptions. Additionally, WS discovery algorithms should perform QoS-based /filtering/ (matchmaking) and /ranking/ (selection) on WS advertisements in order to produce fewer ranked results. A /QoS offer/ (or /demand/) of a WS is a set of constraints on some QoS attributes. Actually, the current modeling efforts of QoS specifications only differ in the expressiveness of these constraints. However, when it comes to QoS attributes modeling, these efforts fail. The main reason is that they use a poor, syntactic and inextensible QoS attributes model. Based on the above problem, the first part of this thesis is dedicated to conducting a research on all QoS-based WS description efforts in order to unveil those features that are necessary for a QoS description model of WSs. The result is a set of requirements for a rich, semantic and extensible QoS-based WS description model. Based on these requirements, an ontology language called /OWL-Q/ has been carefully created by extending OWL-S. One disadvantage of using ontologies is the creation of different descriptions of the same concept. Concerning the QoS domain, this can be true for QoS metrics. We argue that the state-of-the-art ontology alignment algorithms cannot be used in this case because of the use of mathematics for deriving complex QoS metrics from simpler ones. For this reason, we have devised a twofold solution to this problem: a) a semantic QoS metric matching algorithm and b) based on this algorithm, an alignment algorithm for aligning OWL-Q descriptions of WS providers and requesters. There are two main approaches for QoS-based WS matchmaking: ontology-based and /Constraint Programming/ (CP) -based but both of them present some major drawbacks. In this thesis, we argue that the CP-based approach is better but has to be enhanced in many ways. The first enhancement is the use of an appropriate solving technique depending on the type of constraints present in QoS-based WS specifications. The second enhancement is the use of correct matchmaking metrics. Another enhancement is the use of optimization techniques especially for over-constrained demands. All implemented matchmaking algorithms were evaluated based on a testing framework producing random QoS-based WS descriptions. A final contribution of this thesis was a QoS-based WS selection algorithm overcoming the disadvantages of all other research approaches. Thus, as can be seen from the totality of the above contributions, a semantic framework for QoS-based WS description and discovery was produced in terms of this thesis. |
|||
| Language | English | |||
| Issue date | 2008-12-04 | |||
| Collection | School/Department--School of Sciences and Engineering--Department of Computer Science--Doctoral theses | |||
| Type of Work--Doctoral theses | ||||
| Views | 618 | |||
| Digital Documents | |
|---|---|
|
Download document |
