A novel approach to web services discovery
Download
1 / 28

A NOVEL APPROACH TO WEB SERVICES DISCOVERY - PowerPoint PPT Presentation


  • 104 Views
  • Uploaded on

A NOVEL APPROACH TO WEB SERVICES DISCOVERY. Università di Milano-Bicocca. Marco Comerio. [email protected] Index. Objectives & Motivations The Web Services Modeling Ontology (WSMO) Non-functional Properties (NFPs) The NFP-based Discovery Process Conclusions and Future Works.

loader
I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
capcha
Download Presentation

PowerPoint Slideshow about ' A NOVEL APPROACH TO WEB SERVICES DISCOVERY' - kipling


An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript
A novel approach to web services discovery

A NOVEL APPROACH TO WEB SERVICES DISCOVERY

Università di Milano-Bicocca

Marco Comerio

[email protected]

YR-SOC'07 - Leicester 11-12 June 2007


Index
Index

  • Objectives & Motivations

  • The Web Services Modeling Ontology (WSMO)

  • Non-functional Properties (NFPs)

  • The NFP-based Discovery Process

  • Conclusions and Future Works

YR-SOC'07 - Leicester 11-12 June 2007


Objectives
Objectives

  • Definition of a technique to enrich SWS requests and descriptions with the specification of non-functional properties (NFPs);

  • Proposal of a novel approach to a NFP-based SWSs discovery;

  • Test and validation of the proposal in the WSMO environment.

YR-SOC'07 - Leicester 11-12 June 2007


Motivations
Motivations

  • The discovery of a Semantic Web service (SWS) is the act of locating a machine-processable description of a SWS-related resource that may have been previously unknown and that meets certain functional criteria.

  • The increasing availability of SWSs that offer similar functionalities requires the discovery process to be enhanced with a selection phase that considers non-functional properties (NFPs) of the SWSs.

  • NFPs could be quite relevant to match a service request and a service description. In fact, even if a service matches the requested functionalities, it could be unacceptable in terms of NFPs (e.g., cost is too expensive).

YR-SOC'07 - Leicester 11-12 June 2007


Motivations1
Motivations

  • This scenario highlights two needs:

    • enrichment of SWS descriptions with a well-defined set of NFPs;

    • enhancement of the matching functionality between SWS requests and descriptions.

YR-SOC'07 - Leicester 11-12 June 2007


Index1
Index

  • Objectives & Motivations

  • The Web Services Modeling Ontology (WSMO)

  • Non-functional Properties (NFPs)

  • The NFP-based Discovery Process

  • Conclusions and Future Works

YR-SOC'07 - Leicester 11-12 June 2007


The web services modeling ontology wsmo
The Web Services Modeling Ontology (WSMO)

  • WSMO is an ontology for describing various aspects related to Semantic Web Services.

  • WSMO consists of four different main elements:

www.wsmo.org

YR-SOC'07 - Leicester 11-12 June 2007


The web services modeling ontology wsmo1
The Web Services Modeling Ontology (WSMO)

  • Ontologies : provide the terminology;

  • Goals : define the problems to be solved;

  • Web Services : define various aspects of Web Services (e.g., Capabilities, Interfaces);

  • Mediators : allow the overcoming of data, protocol and process heterogeneity. They represent the main advantage of using WSMO and are of four different types: (i) wgMediator; (ii) ggMediator; (iii) ooMediator and (iv) wwMediator.

YR-SOC'07 - Leicester 11-12 June 2007


Index2
Index

  • Objectives & Motivations

  • The Web Services Modeling Ontology (WSMO)

  • Non-functional Properties (NFPs)

  • The NFP-based Discovery Process

  • Conclusions and Future Works

YR-SOC'07 - Leicester 11-12 June 2007


Non functional properties nfps
Non-functional Properties (NFPs)

  • NFPs can be considered to be constraints beyond the functionalities of a Semantic Web Service.

  • The management of NFPs is not a simple task because:

    • NFPs are characterized by several properties;

    • NFPs are of different nature;

    • NFPs are often inter-dependent.

  • Solution: address this problem by developing ontologies of NFPs (OntoNFPs) to formalize definitions, relations, dependencies, heterogeneous measurements and evaluation methods.

YR-SOC'07 - Leicester 11-12 June 2007


Non functional properties nfps1
Non-functional Properties (NFPs)

  • Two different types of NFPs must be considered:

    • offered NFPs (offNFPs) to be included in SWS description;

    • requested NFPs, that are, NFP constraints (conNFPs) to be included in SWS requests.

  • ConNFPs and OffNFPs can be classified as:

    • Qualitative: expressed by enumerative values;

    • Quantitative: expressed by numeric values.

  • The strength of a good offered/required NFP description consists in supporting matching between requested and offered NFPs in a simple and effective way.

YR-SOC'07 - Leicester 11-12 June 2007


Offered nfps offnfps
Offered NFPs (offNFPs)

  • OffNFPs are NFPs supported by the provider of the SWS.

  • The concept of NFP-policy is introduced to collect a set of OffNFPs with a SWS.

  • More NFP-Policies can be associated with a SWS. Each Policy has an applicability condition defined through an axiom.

YR-SOC'07 - Leicester 11-12 June 2007


Nfp constraints connfps
NFP Constraints (conNFPs)

  • ConNFPs are requirements that the requester of a SWS wants to fulfill.

  • The concept of NFP-Request is introduced to collect a set of conNFPs with a Goal.

  • NFP-Request is introduced as a requester-side counterpart of NFP-Policy.

YR-SOC'07 - Leicester 11-12 June 2007


Non functional properties nfps2

Web Service Description

Goal

NFP-Policy

offNFPs

NFP-Request

conNFPs

Non-functional Properties (NFPs)

YR-SOC'07 - Leicester 11-12 June 2007


Index3
Index

  • Objectives & Motivations

  • The Web Services Modeling Ontology (WSMO)

  • Non-functional Properties (NFPs)

  • The NFP-based Discovery Process

  • Conclusions and Future Works

YR-SOC'07 - Leicester 11-12 June 2007


The nfp based discovery process
The NFP-based Discovery Process

  • A novel approach to WS discovery that enriches the traditional UDDI-based process with the use of semantics and evaluation of NFPs.

  • The proposed SWSs discovery process is composed of five different phases.

set up time

execution time

YR-SOC'07 - Leicester 11-12 June 2007


Mediation
Mediation

  • Mediation is performed at setup time and aims at defining:

    • mediators : matching rules for pairs of conNFPs and offNFPs stating when a conNFP is satisfied by an offNFP;

    • meta-mediators : mediators identifying which matching rules need to be invoked for a NFP-Request.

  • Mediators and Meta-mediators are logic programs written in F-logic that are processed by a deductive engine.

  • The result of matching is not a boolean value, it is a positive numeric value that states the degree of satisfaction of a certain constraint.

YR-SOC'07 - Leicester 11-12 June 2007


Matching rules
Matching Rules

  • Matching rules for qualitative and quantitative NFPs have been defined.

  • Quantitative matching rules for the most typical constraint operators have been defined.

  • The matching degree of pairs of quantitative <conNFP, offNFP> depends on the constraint operator op expressed in the NFP constraint and is computed by formulas like

    Degree=fop(OffValue,ReqValue)

YR-SOC'07 - Leicester 11-12 June 2007


Quantitative matching rules examples
Quantitative Matching Rules: Examples

  • Constraint operator (≥ ) : values greater or equal than a reqValue. ReqValue is a lower bound and that the higher value is better.

  • Constraint operator (≥ ) : values at least equal than a reqValue. ReqValue is a lower bound and the lower value is better.

    • if offValue>reqValue

    • if offValue<reqValue

    • if offValue=reqValue

YR-SOC'07 - Leicester 11-12 June 2007


Functional discovery
Functional Discovery

  • Functional Discovery is the traditional SWSs discovery process.

  • This phase has the goal of finding a list of SWSs that totally satisfy the functionalities required by an end-user.

  • This phase finds a set of candidate services; each one providing the same functionality but associated with different sets of policies.

YR-SOC'07 - Leicester 11-12 June 2007


Nfp evaluation
NFP Evaluation

  • NFP Evaluation is in charge of exploiting the matching rules stored in mediators and meta-mediators to compute the matching degree between a conNFP and an offNFP.

  • The result of the NFP Evaluation is a set of values in the range [0..n] associated with each NFP-policy of each SWS identified in the Functional SWS Discovery phase.

YR-SOC'07 - Leicester 11-12 June 2007


Nfp evaluation1
NFP Evaluation

  • On these results, two indicators are defined:

    • normalized degree (n-degree): shows if offValue matches a reqValue and is expressed by values in range [0..1] (if the result is >1 it is reduced to 1).

    • extra matching degree (extraMatch): it is an additional information about how a conNFP is satisfied. It is expressed by values in range [0..1], and is computed by the formula:

YR-SOC'07 - Leicester 11-12 June 2007


Filtering
Filtering

  • Filtering consists in discarding SWSs with NFP-Policies that do not satisfy user minimum requirements and in computing two numeric matching indicator for each NFP-Policy. It is composed of three steps:

    • Mandatory NFPs-driven SWSs discarding : checks if the degree of matching is equal to 1 for all conNFPs with relevance equal to 1. Otherwise, the related NFP-Policy is discharged.

    • Calculus of NFP Global Normalized Value (NGNV) : for each NFP-Policy evaluates the NGNV with the following formula:

    • Calculus of NFP Global ExtraMatch Value (NGEV) : for each NFP-Policy evaluates the NGEV with the following formula:

YR-SOC'07 - Leicester 11-12 June 2007


Ranking
Ranking

  • Ranking is in charge of sorting the policies of the SWSs on the basis of their NGNV values.

  • In the case of policies with the same NGNV value, their NGEV values are considered.

YR-SOC'07 - Leicester 11-12 June 2007


Index4
Index

  • Objectives & Motivations

  • The Web Services Modeling Ontology (WSMO)

  • Non-functional Properties (NFPs)

  • The NFP-based Discovery Process

  • Conclusions and Future Works

YR-SOC'07 - Leicester 11-12 June 2007


Conclusions
Conclusions

  • This ongoing work proposes:

    • an ontological model to capture the nature and relations among NFPs;

    • an enhancement of the discovery process with a selection phase that considers NFPs;

    • two indicators, NGNV and NGEV, to measure the matching degree between requested and offered NFPs, in order to be able to filter, rank and select a set of services offering the same functionality;

    • a set of matching rules to address the different kinds of constraints expressed by the requesters.

YR-SOC'07 - Leicester 11-12 June 2007


Future works
Future Works

  • Experimental validation of the proposed approach in the WSMO environment;

  • Enhance the NFP ontological model to capture the dependency relations among NFPs;

  • Semi-automatic generation of mediators;

  • Management of NFPs in SWS composition.

    These works are supported by the Single European Employment Market Place (SEEMP) project and the Networked Peers for Business (NeP4B) project.

YR-SOC'07 - Leicester 11-12 June 2007


Thank you questions

Thank you!Questions?

YR-SOC'07 - Leicester 11-12 June 2007


ad