Key-based consistency and availability in structured overlay networks

Abstract

Structured Overlay Networks (SONs) provide a promising platform for high performance applications since they are scalable, fault-tolerant and self-managing. SONs provide lookup services that map keys to nodes that can be used as processing or storage resources. In SONs, lookups for a key may return inconsistent results. Consequently, it is difficult to provide consistent data services on top of SONs that build on key-based search. In this paper, we study the frequency of occurrence of inconsistent lookups. We show that the affect of lookup inconsistencies can be reduced by using node responsibilities. We present our results as a trade-off between consistency and availability of keys.

Modeling the Performance of Ring Based DHTs in the Presence of Network Address Translators

“…It is however a well known fact [3,14] that lookup consistency in Chord is compromised when the ring is not perfectly connected. If two nodes queries the system for the same key they should receive identical values but due to node churn a node's pointers might become outdated or wrong before the node had time to correct them in a stabilisation round.…”

Enhanced Paxos Commit for Transactions on DHTs

“…Read and write operations are performed on a majority of replicas, thereby tolerating the unavailability of up to (r − 1)/2 nodes. This scheme is shown to ensure key consistency for data lookups under realistic networking conditions [19].…”

“…Clearly, data consistency cannot be achieved if responsibility consistency is violated. But as shown in [19], the probability of inconsistent data accesses can be reduced by increasing the replication degree, and performing reads on a majority of replicas. In typical Internet scenarios, for example, only three replicas give a consistency probability of five nines.…”

Highly available and scalable grid services

“…The maximum system throughput has been show to scale linearly with the number of nodes. For more information we refer the reader to [8,7]. Scalaris also won the 1st price (shared) at the IEEE Scalable Computing Challenge 2008.…”