Presentation is loading. Please wait.

Presentation is loading. Please wait.

Service Discovery and Semantic Overlay Network Creation in DBGlobe University of Ioannina 4th DBGlobe Meeting Paris, June 23, 2003.

Similar presentations


Presentation on theme: "Service Discovery and Semantic Overlay Network Creation in DBGlobe University of Ioannina 4th DBGlobe Meeting Paris, June 23, 2003."— Presentation transcript:

1 Service Discovery and Semantic Overlay Network Creation in DBGlobe University of Ioannina 4th DBGlobe Meeting Paris, June 23, 2003

2 2Outline zPart 1: Service Discovery zPart 2: Semantic Overlay Networks zPart 3: Ontologies

3 3Outline zPart 1: Service Discovery zPart 2: Semantic Overlay Networks zPart 3: Ontologies

4 4 System Architecture zPMOs are attached to Cell Administration Servers (CASs) zCASs are responsible for the service discovery process zXML data or XML-based service descriptions

5 5 Service Discovery zEach CAS maintains data summaries (e.g. Bloom Filters) to assist query routing A B C SumB SumC

6 6 Multi-level Bloom Filters zHash-based indices that extend Bloom filters to support the evaluation of path queries. zTwo approaches: Breadth and Depth Bloom Filters that rely on different ways of hashing an XML-tree. zCompact structures zAppearance of false positives

7 7 Performance Results zMulti-level Bloom filters outperform in terms of false postives Simple Bloom filters in evaluating path queries. zFor 2% of the total size of the data, multi- level Bloom filters evaluate path queries for a false positives ratio below 3% zBreadth Blooms work better than Depth Blooms. zDepth Blooms require more space but are suitable for special type of queries.

8 8 Filters Distribution zPeers organized into hierarchies connected through a main channel zEach server maintains: ya local filter ya merged filter of the filters in its sub-tree. yIf it is a root-peer(connected to the main channel) a merged filter for every other root-peer

9 9 Distribution: Hierarchical Organization Node C: Local filter Merged filter :E  F  G  H Root filters: A, B, D

10 10Outline zPart 1: Service Discovery zPart 2: Semantic Overlay Networks zPart 3: Ontologies

11 11 Content-based organization zGroup peers together according to their content zUse filter and not data similarity for efficiency zWhen a peer joins the system: it broadcasts its local summary and attaches to the most «similar» peer available

12 12 Bloom Filter Similarity zNodes organized according to Bloom Filter Similarity zMeasure: similarity measure based on the Manhattan distance metric. Let two filters B and C of size m d(B, C) = |B[1] – C[1]| + |B[2] – C[2]| + … |B[m] – C[m]|. similarity(B, C) = m – d(B, C).

13 13 Bloom Filter Similarity (cont’d) 1 011 0 001 110 0 100 1 C B similarity(B, C) =8 - (0 + 1 + 1 + 0 + 1 + 0 + 1 + 0) = 4 For multi-level Bloom filters similarity is defined as the sum of each pair of corresponding levels

14 14 Performance Results zThe content-based organization is much more efficient in finding all the results for a query, than the proximity organization. z They both perform similarly in discovering the first result. z The content-based organization outperforms the proximity one when the nodes that satisfy a given query are limited.

15 15 Current work zA peer can belong to more than one hierarchies. zSelf-organization--tuning of the system predefined threshold. zService Discovery: yLocate the right cluster (hierarchy) yFind the peer in the hierarchy

16 16Outline zPart 1: Service Discovery zPart 2: Semantic Overlay Networks zPart 3: Ontologies

17 17 Ontologies Ontologies are hierarchies --> Thus they can be summarized by multi-level Blooms

18 18 Ontologies (cont’d) zThe main issue: How to locate the matching hierarchy(cluster) zJust check every root peer. Can we use a global ontology to route us to the matching hierarchy more efficiently?

19 19 Thank you


Download ppt "Service Discovery and Semantic Overlay Network Creation in DBGlobe University of Ioannina 4th DBGlobe Meeting Paris, June 23, 2003."

Similar presentations


Ads by Google