1. A service-oriented middleware for building context-aware services Center for E-Business Technology Seoul National University Seoul, Korea Tao Gu, Hung Keng Pung, Da Qing Zhang Institute for Infocomm Research, Singapore Journal of Network and Computer Applications, 2005 Presented by Nam, Kwang Hyun Intelligent Database Systems Lab School of Computer Science & Engineering Seoul National University, Seoul, Korea

2. Copyright  2008 by CEBT Contents  Introduction  Context modeling and reasoning  The SOCAM architecture  Performance evaluation  Conclusion 2

3. Copyright  2008 by CEBT Introduction  Context-aware service A network service which uses various contexts and adapts itself to the change of environment dynamically and automatically  Architecture requirements A common context model that can be shared by all devices and serv ices A set of services that perform – Context acquisition – Context discovery – Context interpretation – Context dissemination 3

4. Copyright  2008 by CEBT Proposal  Model An ontology-based context model using OWL  System A Service-Oriented Context-Aware Middleware (SOCAM) – Includes a set of independent services. – Supports Acquiring various contexts from different context providers Interpreting contexts through context reasoning Delivering contexts in both push and pull modes. 4

5. Copyright  2008 by CEBT Context Model  Ontology a vocabulary for representing knowledge about a domain and for descri bing specific situations in a domain  An ontology-based approach Allows to describe contexts semantically in a way which is independent of programming language, underlying operating system or middleware  Context reasoning using first-order logic, temporal logic, and others enables to be done  Contexts are represented as first-order predicate calculus Predicate(subject, value) – Location(John, bathroom) – Temperature(kitchen, 120) 5

6. Copyright  2008 by CEBT Context Ontology  Two-layer hierarchical approach for designing context ontologies. Common upper ontology – For the general concepts Domain-specific ontologies – Apply to different sub-domains  Benefit of two-layer hierarchical approach Reduces the scale of context knowledge Releases the burden of context processing for pervasive devices in each domain 6 의코 인논

7. Copyright  2008 by CEBT Upper(Generalized) Ontology 7  Basic concepts Person Location Computational entity & activity  Composition 14 classes 6 properties

8. Copyright  2008 by CEBT Domain-Specific ontology  Domain-specific ontol ogy defines The details of gener al concepts Their properties 8 Example subClassOf IndoorSpace Room Corridor Entry

9. Copyright  2008 by CEBT Context classification  Direct context Directly acquired or obtained from a context provider Sensed context – Acquired from physical sensors (e.g. door’s status) Defined context – Defined by a user (e.g. user’s foodPreference)  Indirect context Derived by interpreting direct context through context reasoning Example – Showing can be inferred from Bathroom, (Water heater) On, (Door) Clos ed  Provide an additional property elements – owl:classifiedAs 9

10. Copyright  2008 by CEBT Class dependency  Dependency Captures the existence of a reliance of property associated with one entity on another. Provide an additional property elements – rdfs:dependsOn  The importance of context dependency Enable to incorporate probability and Bayesian networks to reason a bout uncertain contexts. 10

11. Copyright  2008 by CEBT The SOCAM architecture  A distributed middleware that transfers and converts various phys ical spaces from which contexts are acquired into a semantic spa ce where contexts can be easily shared and accessed by context -aware services 11

12. Copyright  2008 by CEBT The SOCAM architecture’s components  Context provider Abstract useful contexts from heterogeneous sources Convert them to OWL representation to share and reuse contexts External (from external source) and Internal (from ubiquitous sensors)  Context interpreter Provides logic reasoning services to process context information Context reasoner – Provide deduced contexts – Detect inconsistency and conflict in context KB Context KB – Provide a set of API’s for other service components to query, add, delete or modify context knowledge 12

13. Copyright  2008 by CEBT The SOCAM architecture’s components  Context database Stores context ontologies and past contexts for a sub-domain  Service locating service Provides a mechanism where context providers and context interpre ter can advertise their presence  Context-aware services Make use of different level of contexts Adapt the way they behave according to the current context 13

14. Copyright  2008 by CEBT Implementation  SOCAM middleware implemented in J2SE 1.3.1  Context interpreter implemented using Jena2-HP’s Semantic We b Toolkit  Domain Specific ontologies implemented in OWL  Home domain ontology 89 classes 156 properties  Vehicle domain ontology 32 classes 57 properties 14

15. Copyright  2008 by CEBT Performance Evaluation (1/4) 15 Overhead of the two-layer ontology design

16. Copyright  2008 by CEBT Performance Evaluation (2/4) 16 The reasoning performance

17. Copyright  2008 by CEBT Performance Evaluation (3/4) 17 Reasoning comparison

18. Copyright  2008 by CEBT Performance Evaluation (4/4) 18 Average time for concurrent requests

19. Copyright  2008 by CEBT Conclusion  Present a formal context model based on OWL  The SOCAM middleware has been designed to support the buildi ng of context-aware services  The evaluation results demonstrate a reasonable performance  It is able to meet the requirements of context-aware systems 19

20. Copyright  2008 by CEBT Discussion  Pros Considering context dependency is novel. Independent service components enables this architecture to operat e in distributed and heterogeneous networks.  Cons Performance is really reasonable? 20