1 / 15

A P2P Semantic-Based Service Discovery Method for Pervasive Computing Environment

A P2P Semantic-Based Service Discovery Method for Pervasive Computing Environment. 2007.10.31 Shin Yongjin. Contents. Introduction Multi-level Structure Service Discovery Algorithm Contact Selection Mechanism Evaluation Conclusion. Introduction. Peer to Peer Search

micheal
Download Presentation

A P2P Semantic-Based Service Discovery Method for Pervasive Computing Environment

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. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. A P2P Semantic-Based Service Discovery Method for Pervasive Computing Environment 2007.10.31 Shin Yongjin

  2. Contents • Introduction • Multi-level Structure • Service Discovery Algorithm • Contact Selection Mechanism • Evaluation • Conclusion

  3. Introduction • Peer to Peer Search • Unstructured: flooding queries to all peers, use data replication algorithm. • Structured: using DHT • Semantic-based service discovery scheme • A portion of service attributes are registered with index • Service attributes are hierarchically organized. • Small world network

  4. Multi-level Structure Hierarchical Service Attribute Tree Function Class ( ) Location Information ( ) Invoking Interface( ) Regulation Parameter( ) Priority :

  5. Multi-level Structure • Service Discovery Network • Three Level of topology • Node level topology: physical communication relation among nodes • Location level topology: location information about the physical space • Service level topology: relation expression of service function class information  Enhance semantic connection degree of node.  Reduce the search space in order for lesser discovery time and lesser node disturbance.

  6. Service Discovery Algorithm • Service function class = service key of node • Local service function class – primary key • Non-local service function class – secondary key • Every node has two node sets. • Small world network • Distant nodes = contact • Local contact with small average length. • Small number of long range contacts with large clustering coefficient.

  7. Service Discovery Algorithm • Hybrid of proactive and reactive • Advertisement diameter is determined by node capability and mobility possibility • When receiving advertisement message, node can modify advertisement hop counts according to own capability • Node can specify the predicate of advertisement and forward

  8. Contact Selection Mechanism • Kindred-Based Contact Selection • Two considerations • Find the long contact node containing more service function class information and overlapping fewer dominant community of service request node • Find a long contact for establishing the kindred node topology to guarantee the higher reachability

  9. Contact Selection Mechanism • Three Steps • Frontier Node (FN) Selection From the center q, it starts finding FN. At first, q visits one node, and its tag becomes true, and it is set to FN. Visit next and changes tag to true and check the FN or not. u q_f FN, tag=true; FN, tag=true; d q FN, tag=true; tag=true; tag=true;

  10. Contact Selection Mechanism • Contact Discover Direction • Finding exterior neighbor node(EN) u and c are not belong to service community of q To forward contact discover message With EN, can select long contact discover direction for avoiding the unstable path. u c EN u EN u FN q_f FN u_f q

  11. Contact Selection Mechanism • Contact Node Selection • If long contact node is kindred node of query node, then, directly select this node to be long contact node. • If not, • Similarity check with RSI • RSI (Relative Semantic Information value) • Represents similarity between two nodes. • Ratio of the number of same service function class and total number of contact node.

  12. Contact Selection Mechanism • Location-Aware Contact Selection • Location level topology • Choose the node with different location information from service community of query node. • Optimal average hop in querying the service with specified location requirement.

  13. Evaluation • Balance capacity of discovery efficiency • Average overhead in discovering different popularity service • Predict the overhead proportion relation of service discovery method • Discover a service with different popularity = = Message overhead = average effect of discovery result = function about similarity between different service function class discovery

  14. Evaluation • For simulation, • Use 10 services • J-Sim Simulator • Service advertisement diameter = 3

  15. Conclusion • Contact based small world network • Low maintenance overhead • Efficient search performance • My opinion • Only for small size network. • It takes too much time for finding FN and EN • So, it is not appropriate to big size network.

More Related