Hybrid Approach to Collaborative Context-Aware Service Platform for Pervasive Computing Dejene Ejigu, Marian Scuturici, Lionel Brunie Laboratoire LIRIS-UMR-CNRS 5205, INSA de Lyon,France JOURNAL OF COMPUTERS, VOL. 3, NO. 1, JANUARY Presented by Yeon JongHeum, IDS Lab., Seoul National University

Copyright  2008 by CEBT Introduction  Enhanced CoCA : a Collaborative Context-Aware service platform  HCoM : CoCA is based on Hybrid Context Management model  Platform : Neighborhood collaboration mechanism  Initial prototype of the platform Good standard of scalability 2

Copyright  2008 by CEBT Conceptual Framework for Pervasive Context-Awareness  Basic elements of a pervasive computing environment 3

Copyright  2008 by CEBT Pervasive Environment  Characteristics Dynamicity Heterogeneity and Ubiquity – Users / Devices / Resources Ad-hoc connection among the devices Existence of hardware and software sensors  Plenty of interacting gadgets tiny and invisible  Users in are sometimes frustrated by the gadgets  Challenge to know when and under what situation they operate these gadgets and the time allocated for such activity 4

Copyright  2008 by CEBT Context Modeling  HCoM modeling deals with how context data is … Collected Organized Represented Stored Presented 5

Copyright  2008 by CEBT Collaborative Context-Aware Service  CoCA service Interprets and Aggregates context value Performs Reasoning about context Passes decisions about the actions to be triggered Stores knowledge and decisions  Necessary of “Collaboration” and “Combined Efforts” of Devices Devices are tiny and resource hungry Context-aware service requires a large amount of hardware resources 6

Copyright  2008 by CEBT Related Works  Application development / few types of context information : identity, time, location Cyber Guides Cool Town Cricket Compass  Major contributions / does not allow ad-hoc communications Oxygen Gaia  Semantic Web Technology CoBrA-ONT  Context Ontology CONON  Service based framework / Middle ware solution A Service-Oriented Middleware for Building Context-Aware Services  Emphasized the importance of ad-hoc communication MARKS RCSM 7

Copyright  2008 by CEBT Related Works  CoCA: A Collaborative Context-Aware Service Platform for Pervasive Computing  An Ontology-Based Approach to Context Modeling and Reasoning in Pervasive Computing  Hybrid Approach to Collaborative Context-Aware. Service Platform for Pervasive Computing.  Semantic approach to context management and reasoning in ubiquitous context-aware systems 8

Copyright  2008 by CEBT Reverse Salient & Proposal  Bottleneck for the development of pervasive applications Computationally intensive characteristics of context reasoning process Lack of semantically rich context model  Neighborhood based collaborative context-aware service platform facilitates reusability of context resources and reasoning axioms sharing of computational resources among devices in the neighborhood space using semantic ontology based hybrid model as a core data source 9

Copyright  2008 by CEBT Model : Semantic Based Hybrid Context Model  Context can be described statement of characteristics of a context entity Transitive – e.g. “library is located in campus” & “student is located in library” → “student is also located in campus”. Inverse – e.g. “device is owned-by person” → “person owns device” Time / Precision – e.g. “student is located in library” at a given time t – e.g. “a network connection has low speed” is 85% etc. 10 Generic levelDomain level equivalent Person isEngagedIn ActivityPhysician isEngageIn Patient-treatment Person isLocatedIn LocationStudent isLocatedIn Library

Copyright  2008 by CEBT Model : Context Representation  EHRAM Hierarchies (H) of set of Entities (E) / set of entity Relations (Re) / set of attribute Relations (Ra) Set of axioms (A) / set of metadata (M) 11 Domain Layer Generic Layer

Copyright  2008 by CEBT Model : Context Representation (cont’d)  RDF-triple :  Reification Additional metadata about the basic context triple can also be included as part of the context data e.g. “Bob is located in the Library”, “is reported by sensor #5”, “has accuracy of 88%” 12 //Reification //original statement //reification starts //reif. ends // meta-context on XX // meta context on XX

Copyright  2008 by CEBT Model : Context Representation (cont’d)  EHRAM model Representation of semantics of the data (context knowledge) – Using Semantic Ontology – OWL and RDF schema Representation of the context data – Mapping into Relation Database – Based on ER to EHRAM model mapping algorithm 13

Copyright  2008 by CEBT Model : Hybrid Approach to Context Management  Hybrid Context Management (HCoM) Ontology approach to manage context semantics Relational approach to manage context data. Process them separately and put the results together for better reasoning and decision support in a context-aware environment Upgrade on GCoM(Generic Context Management model)  Heuristic component on HCoM Loading only relevant data for reasoning – Minimizes the size of the reasoning space – Reduces the unnecessary overloading of the reasoner Improve the overall performance Overcome limitations of lack of scalability 14

Copyright  2008 by CEBT Model : Hybrid Approach to Context Management  Architecture of HCoM Model 15 requirement and preference for heuristic selection query and retrieval of relevant static context data during initialization relevant static context data populated into the ontology

Copyright  2008 by CEBT Architecture for the CoCA platform  Interface Manager Controls UI and an Interface between the CoCA platform and other modules Hosts action triggering process  Data Source a heuristic component to handle context selection and loading from relational data store to the semantic schema to facilitate speedy reasoning  Core Service provides the context-aware service after performing reasoning and decisions  Supplementary Service Knowledge discovery & Collaboration service 16

Copyright  2008 by CEBT Component view of the Enhanced CoCA platform  The RAID Action Engine  Context data User interface Sensors  Rules User interface Rule mining module  Ontology Generic ontology Domain dependant ontology 17

Copyright  2008 by CEBT CoCA Action Triggers  Principles of multifaceted action processing in CoCA  Application consists of both Components context sensitive Components that do not depend on the context A context sensitive component can be seen as an item with many facets  In CoCA platform, a facet has a condition that behaves like a switch in the sense that if the condition is true the facet is exposed otherwise the facet is hidden 18

Copyright  2008 by CEBT Collaboration Manager  Share computing resources context / rules / ontology / processor / memory / etc.  JXTA as a supporting technology is a set of open protocols Allow any connected device on the network ranging from cell phones and wireless PDAs to PCs, servers and super computers Communicate and collaborate in a peer-to-peer manner 19

Copyright  2008 by CEBT Activity diagram for the CoCA service platform 20 units responsible to do the activity (the base unit itself or the neighborhood space) services responsible to do the activity (the context modeling service or the context-aware service)

Copyright  2008 by CEBT The PiCASO Scenario  Scenario – Pervasive Campus Aware Smart Onlooker example (PiCASO) Scenario of a university campus where research students and professors are involved Regular meetings, informal and spontaneous meetings and discussions are important for the advancement of their work Discussion can take place among two or more depending on the relevance of their work 21

Copyright  2008 by CEBT Evaluation  Scalability Measures for GCoM and HCoM Only relevant context is loaded for reasoning at a time 22

Copyright  2008 by CEBT Conclusion & Discussion  Pros Hybrid Approach to solve scalability of ontology  Cons Just combined existed technique  Discussion What’s Next? 23