Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.491946
Title: Efficient access to distributed information using structured peer-to-peer systems
Author: Gonzalez Beltran, Alejandra Noemi
Awarding Body: Queen's University of Belfast
Current Institution: Queen's University Belfast
Date of Award: 2008
Availability of Full Text:
Full text unavailable from EThOS. Please contact the current institution’s library for further details.
Abstract:
This thesis examines the problem of accessing distributed information about resources in large-scale, dynamic and heterogeneous computing environments. A mechanism with this functionality is an information service and facilitates the discovery of available resources. Grid computing and Peer-to-Peer (P2P) computing are resourcesharing environments as above, where ap. information service is essential. Structured P2P systems offer suitable routing substrates for information services. They organise nodes in a network providing a scalable, fully-decentralised and selforganising model for routing queries among the nodes. In addition, they support bounded searches with guaranteed results. However, they exhibit some limitations. First, they assume homogeneous nodes, simplifying the environment characteristics and limiting the performance and scalability of the network. Second, their querying capabilities are limited, not fully supporting multi-dimensional range and aggregation queries as required by an information service. In order to overcome these limitations, this thesis presents the following contributions. Firstly, the homogeneity assumption is removed. The required components for a structured P2P-based information service are.identified and heterogeneity-aware techniques are proposed for each component. These techniques involve improvement of routing latency and online load balancing algorithms in heterogeneous settings. Secondly, improvements over the routing substrate are considered. As a result, . a new structured P2P network called skip tree graph is proposed. Skip tree graph is the first system simultaneously supporting range queries, aggregation queries and multicasting/broadcasting facilities. It constitutes a hierarchical and randomised distributed data structure that allows performance improvements in all the querying algorithms. In particular, several exact-match and range query schemes over skip tree graphs are introduced and compared. An efficient repair mechanism for skip tree graph is described, which makes the structure resilient to failures. The contributions are examined following both theoretical and experimental approaches. The experimental validation is based on event-driven simulations.
Supervisor: Not available Sponsor: Not available
Qualification Name: Queen's University of Belfast, 2008 Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.491946  DOI: Not available
Share: