Declarative programming over eventually consistent data stores

Abstract

User-facing online services utilize geo-distributed data stores to minimize latency and tolerate partial failures, with the intention of providing a fast, always-on experience. However, geo-distribution does not come for free; application developers have to contend with weak consistency behaviors, and the lack of abstractions to composably construct high-level replicated data types, necessitating the need for complex application logic and invariably exposing inconsistencies to the user. Some commercial distributed data stores and several academic proposals provide a lattice of consistency levels, with stronger consistency guarantees incurring increased latency and throughput costs. However, correctly assigning the right consistency level for an operation requires subtle reasoning and is often an error-prone task. In this paper, we present QUELEA, a declarative programming model for eventually consistent data stores (ECDS), equipped with a contract language, capable of specifying fine-grained applicationlevel consistency properties. A contract enforcement system analyses contracts, and automatically generates the appropriate consistency protocol for the method protected by the contract. We describe an implementation of QUELEA on top of an off-the-shelf ECDS that provides support for coordination-free transactions. Several benchmarks including two large web applications, illustrate the effectiveness of our approach.

CLOTHO: directed test generation for weakly consistent database systems

“…Although strictly serial executions are easily comprehensible and desirable, forcing the database to execute transactions one-by-one is unnecessarily restrictive and must usually rely on poorly scalable pessimistic locking mechanisms to enforce [Sivaramakrishnan et al 2015]. We instead define an execution history H to be serializable if there exists another strictly serial history H ′ , that is constructed by reordering execution steps of H , such that the final set of effects in both histories are equivalent.…”

Characterizing and Adapting the Consistency-Latency Tradeoff in Distributed Key-Value Stores

“…Recently, there has been work on declarative ways to specify application consistency and latency requirements -PCAP proposes mechanisms to satisfy such specifications [55].…”

Sequential programming for replicated data stores

“…We described how our work builds on CRDTs (Shapiro et al [Shapiro et al 2011] provide a comprehensive overview). Several frameworks allow both conflict-free, and conflicting operations [Balegas et al 2015;Li et al 2014Li et al , 2012Sivaramakrishnan et al 2015;Terry et al 1995], offering different trade-offs between consistency and availability. Such mixed-consistency systems are typically built upon key-value databases that offer tunable transaction isolation [Bailis et al 2013;Lakshman and Malik 2010;Terry et al 2013].…”

“…A related approach [Li et al 2012;Sivaramakrishnan et al 2015] allows manual selection of consistency levels for operations. Quelea [Sivaramakrishnan et al 2015] allows specifying contracts (ordering constraints) on effects. In contrast, our system hides the concept of effect ordering in history, and allows modular conflict specification.…”

“…Or more precisely, a division of the application into strong and eventual components. Mixed consistency programming systems designed to streamline this division task include RedBlue [Li et al 2012], in which programmers specify łwithdraw→Redž or łdeposit→Bluež to set strong or eventual behavior for individual operations, and Quelea [Sivaramakrishnan et al 2015], in which programmers write explicit łwithdraw × withdrawž mutex annotations that the runtime system will respect.…”