Recent work on peer-to-peer systems has demonstrated the ability to deliver low latencies and good load balance when demand for data items is relatively uniform. We describe a lightweight, adaptive, and system-neutral replication protocol, LAR, that delivers low latencies and good load balance even when demand is heavily skewed.Simulation of LAR in combination with both the Chord and TerraDir systems shows that LAR quickly adapts to non-uniformity in both the underlying system topology and in the input stream. Further, we demonstrate better performance than functionally similar application-layer protocols, using an order of magnitude less network bandwidth.
We describe d-spaces, a replica contml protocol dejined in terms of quorum sets on multi-dimensional logical structures. Our work is motivated by asymmetrical access patterns, where the number ofread accesses to data are dominant relative to update accesses, i.e. where the protocols should be read-few write-many. D-spaces are optimal with respect fa quorum group sizes. The qrraliry of the tradeoff between read efficiency and update availability is not matched by existing quorum protocols. We also propose a novel scheme for implementing d-spaces that combines caching and local information to provide a best-effort form of global views. This allows quorum reconJiguration to be lightweight without impacting access latencies, even when the rate of membership changes is very high.
We present the design and evaluation of the query-routing protocol of the TerraDir distributed directory. TerraDir is a wide-area distributed directory designed for hierarchical namespaces, and provides a lookup service for mapping keys to objects. We introduce distributed lookup and caching algorithms that leverage the underlying data hierarchy. Our algorithms provide efficient lookups while avoiding the load imbalances often associated with hierarchical systems. The TerraDir load balancing scheme also incorporates a node replication algorithm that provides configurable failure resilience with provably low overheads.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.