Presentation is loading. Please wait.

Presentation is loading. Please wait.

University of Maryland, Baltimore County

Similar presentations


Presentation on theme: "University of Maryland, Baltimore County"— Presentation transcript:

1 University of Maryland, Baltimore County
DReggie: Semantic Service Discovery for M-Commerce Applications Dipanjan Chakraborty Filip Perich Sasikanth Avancha Anupam Joshi

2 Today: Life is easy

3 Tomorrow: We Got Problems

4 Service Discovery Techniques
Unique ID based Attribute-based Interface-based Some existing discovery architectures Jini Salutation SLP UPnP UDDI Bluetooth SDP

5 NO Is this enough? Lack of rich representation
Lack of constraint specification and inexact matching Lack of ontology support Resource considerations for mobile devices

6 DReggie: A semantic Service Discovery System
Built on Jini Technology Uses DARPA Agent Markup Language (DAML) to describe services Services matched using attribute, interface and their semantic description Simple Java-based light-weight reasoner and a complex Prolog-based reasoner Inexact matching, resource-based matching

7 DReggie Architecture Mobile Device DAML Register Service DAML
Lookup Server Service Description Matching Module DAML Service Register DAML Request Mobile Device Reply Reply Reply Register DAML Service

8 Ontology Details Use resource description capabilities of DAML+OIL to represent services and request An Ontology to represent services in terms of Capabilities Input/Outputs Platform dependencies Current resources Mobility

9 Java-based simple reasoning module for lightweight devices
Lightweight Matching Java-based simple reasoning module for lightweight devices Module parses the DAML request Tries to match the ‘nearest’ matching service Takes into consideration resource limitations and dependencies of services

10 Heavyweight Matching Heavyweight Prolog-based reasoning module for resource-rich devices Parses the DAML ontology, service profile and loads the facts into its knowledge base Parses the request and uses inference techniques to match for services Capability to match services similar to the one requested

11 DAML+OIL (DARPA Agent Markup Language +Ontology Inference Layer)
Why DAML+OIL? DAML+OIL (DARPA Agent Markup Language +Ontology Inference Layer) Building tool for the semantic web Rich description capability to describe resources/services Incorporation of simple rules Built on top of RDF/XML

12 Limitations/Future Directions
Directory-based service discovery Peer-to-peer discovery? Ad-hoc environment Network/language dependence Future Directions Richer matching capabilities to the Prolog reasoner Peer-to-peer discovery in ad-hoc environments

13 Ebiquity Research Group at UMBC
Resources Service Ontology Ebiquity Research Group at UMBC

14 Ebiquity Research Group


Download ppt "University of Maryland, Baltimore County"

Similar presentations


Ads by Google