2011 | OriginalPaper | Buchkapitel
Removing Uncertainties from Overlay Network
verfasst von : Ye Yuan, Deke Guo, Guoren Wang, Lei Chen
Erschienen in: Database Systems for Advanced Applications
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Overlay networks are widely used for peer-to-peer (P2P) systems and data center systems (cloud system). P2P and data center systems are in face of node frequently joining, leaving and failure, which leads to topological uncertainty and data uncertainty. Topological uncertainty refers to that overlay network is incomplete, i.e., failures of node and link (between two nodes). Data uncertainty refers to data inconsistency and inaccurate data placement. Existing P2P and data center systems have these two uncertainties, and uncertainties have an impact on querying latency. In this study, therefore, we first give probabilistic lower bounds of diameter and average query distance for overlay network in face of these two uncertainties. The querying latency of existing systems cannot be better than the bounds. Also, existing systems often suffer unsuccessful queries due to uncertainties. To support an efficient and accurate query, we propose a topology constructive method and a data placement strategy for removing two uncertainties from overlay network. Also, efficient algorithms are proposed to support range queries in an overlay network. The DeBruijn graph representing overlay network is used to construct a new system,
Phoenix
, based on proposed methods. Finally, experiments show that performances of
Phoenix
can exceed the probabilistic bounds, and they behave better than existing systems in terms of querying latency, querying costs, fault tolerance and maintenance overhead.