An efficient scheme for providing high availability

Bhide, Anupam; Goyal, A.; Hsiao, Hui-I; Jhingran, Anant

doi:10.1145/130283.130320

Cited by 10 publications

(3 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several research projects and commercial products use replication to avoid the downtime caused by crashes [3,4,12,14,9,5]. Systems can typically be classified into eager or lazy replication based on whether they propagate updates inside transaction boundaries or not.…”

Section: Database Replication For Crash Faultsmentioning

confidence: 99%

“…Systems can typically be classified into eager or lazy replication based on whether they propagate updates inside transaction boundaries or not. They also differ in whether the clients observe a single-copy serializable view [14,5,4] or a more lax concurrency model [20,12,10]). Some work also explores the problem of providing availability in a partitioned system [1].…”

Section: Database Replication For Crash Faultsmentioning

confidence: 99%

See 1 more Smart Citation

Tolerating byzantine faults in transaction processing systems using commit barrier scheduling

Vandiver

Balakrishnan

Liskov

et al. 2007

Proceedings of Twenty-First ACM SIGOPS Symposium on Operating Systems Principles

View full text Add to dashboard Cite

This paper describes the design, implementation, and evaluation of a replication scheme to handle Byzantine faults in transaction processing database systems. The scheme compares answers from queries and updates on multiple replicas which are unmodified, off-the-shelf systems, to provide a single database that is Byzantine fault tolerant. The scheme works when the replicas are homogeneous, but it also allows heterogeneous replication in which replicas come from different vendors. Heterogeneous replicas reduce the impact of bugs and security compromises because they are implemented independently and are thus less likely to suffer correlated failures.The main challenge in designing a replication scheme for transaction processing systems is ensuring that the different replicas execute transactions in equivalent serial orders while allowing a high degree of concurrency. Our scheme meets this goal using a novel concurrency control protocol, commit barrier scheduling (CBS). We have implemented CBS in the context of a replicated SQL database, HRDB (Heterogeneous Replicated DB), which has been tested with unmodified production versions of several commercial and open source databases as replicas. Our experiments show an HRDB configuration that can tolerate one faulty replica has only a modest performance overhead (about 17% for the TPC-C benchmark). HRDB successfully masks several Byzantine faults observed in practice and we have used it to find a new bug in MySQL.

show abstract

Section: Database Replication For Crash Faultsmentioning

confidence: 99%

Section: Database Replication For Crash Faultsmentioning

confidence: 99%

Tolerating byzantine faults in transaction processing systems using commit barrier scheduling

Vandiver

Balakrishnan

Liskov

et al. 2007

Proceedings of Twenty-First ACM SIGOPS Symposium on Operating Systems Principles

View full text Add to dashboard Cite

show abstract

“…Writeahead logging schemes such as Algorithm for Recovery and Isolation Exploiting Semantics (ARIES) [24] bring the system to the latest transaction-consistent state after a failure. Utilizing asynchronous replication to create secondary logbased replicas for increasing availability in such systems has been proposed by Bhide et al [8]. The approach of such systems is to maintain the secondary to a consistent point with the primary as recent in time as possible without regard to the recovery value of individual objects within the replica.…”

Section: Related Workmentioning

confidence: 99%

Using adaptive scheduling for increased resiliency in passive asynchronous replication

Adams

Gračanin

2007

Innovations Syst Softw Eng

View full text Add to dashboard Cite

Distributed systems commonly replicate data to enhance system dependability. In such systems, a logical update on a data item results in a physical update on a number of copies. The synchronization and communication required to keep the copies of replicated data consistent introduce a delay when operations are performed. In systems distributed over a bandwidth-constrained area, such operational delays generally prove unacceptable and passive asynchronous replication is often used to mitigate the delays. The research described in this paper looks to develop a new methodology for passive asynchronous replication that includes the introduction of an adaptive data replication scheduler to increase the data redundancy of the most valued data objects in overloaded situations. The approach we describe relies on the batch processing nature of passive asynchronous replication to make update decisions based on a defined policy. The methodology uses an adaptive scheduling algorithm providing near real-time selection of which objects to replicate during overloaded conditions based on a trained multilayer perceptron neural network. Historical replication logs are used for the initial training of the network and supervised training continues as replication logs are created providing periodic improvement through a feedback mechanism. This paper presents summary results of data replication scheduling simulations with an emphasis on the design selection of the adaptive distributed scheduler.

show abstract