a context information service using ontology-based queries ruaidhrí power, dave lewis, declan...
TRANSCRIPT
A Context Information Service using Ontology-Based Queries
Ruaidhrí Power, Dave Lewis, Declan O’Sullivan, Owen Conlan, Vincent Wade
Knowledge and Data Engineering GroupDepartment of Computer Science, Trinity College Dublin
Context Information Service
● This paper describes the design of a context information service that serves ontology-based context queries– Motivation– Architecture– Query Interface– Query Routing– Scalability
Architecture Motivation
● Current approaches advocate predefined, agreed models for context information– Context information cannot be globally standardized
● Cannot be certain what information is relevant before constructing the system– Need capability to update models
● Must cope well with heterogeneous systems– Translate queries (and responses) into terms in each
system’s own ontology
Context Self-Management
● Context will be managed by each user’s own systems, rather than centrally– Global view of all context impossible– Privacy and security concerns
● Interoperability between heterogeneous context sources
● External context information will be merged into each user’s own view of the world, through the use of ontologies
Context Service Node
● A Context Service Node (CSN) receives context queries and resolves them by acting as mediator to other context services
● Translate query responses into terms in each application’s own ontology
● Also accepts registration of ontologies and mappings between ontologies
● Ontologies and mappings can be provided separately, and updated as needed
Context Querying
● CSN query interface– SQL, XQuery, RDQL, ...
● Application poses queries using terms from its own registered ontologies
● CSN communicates with other CSNs over the network to resolve query
● Query translation and routing are based on ontological information available
Context Service Network Architecture
Application seeking context
CSN
Context Service Network
Reg(O1,O2) Query(Q1,O1) Resp(R)
Context Service Network Architecture
Application seeking context
CSN
Context Service Network
Reg(O1,O2) Query(Q1,O1) Resp(R)
Ontology mapping
respository Map (O1,O3)
MapReq(O1)
Context Service Network Architecture
Application seeking context
CSN
CSN CSN
CSN
CSN
Context Service Network
Reg(O1,O2) Query(Q1,O1) Resp(R)
Ontology mapping
respository Map (O1,O3)
MapReq(O1)
Context Service Network Architecture
Application seeking context
Context info
source
CSN
CSN CSN
CSN
CSN
Context info
source
Context Service Network
Reg(O3)
Reg(O1,O2) Query(Q1,O1) Resp(R)
Query(Q1,O3) Resp(O3)
Ontology mapping
respository Map (O1,O3)
MapReq(O1)
Routing Strategies
● Broadcast inefficient except in small networks● Must route queries to nodes that understand them● What if a large number of CSNs understand a
query?– Trade off number of nodes vs. processing cost
Content-Based Networking (CBN)
● Messages routed based on their content● Content-based routers route messages using
function which describes which messages their neighbours are interested in
● Persistent ontology-based queries (pub/sub)● CBN systems have shown promising scalability
Conclusion
● Domain knowledge of applications encapsulated in an ontology
● Ontologies used as the basis for interoperability● System integration performed using mappings, as
a separate activity● Intelligent pub/sub (CBN) for query routing● OWL as ontology language, XQuery as query
language