Quantifying isolation anomalies

Abstract

Choosing a weak isolation level such as Read Committed is understood as a trade-off, where less isolation means that higher performance is gained but there is an increased possibility that data integrity will be lost. Previously, one side of this trade-off has been carefully studied quantitativelythere are well-known metrics for performance such as transactions per minute, standardized benchmarks that measure these in a controlled way, and analytic models that can predict how performance is influenced by system parameters like multiprogramming level. This paper contributes to quantifying the other aspect of the trade-off. We define a novel microbenchmark that measures how rapidly integrity violations are produced at different isolation levels, for a simple set of transactions. We explore how this rate is impacted by configuration factors such as multiprogramming level, or contention frequency. For the isolation levels in multi-version platforms (Snapshot Isolation and the multiversion variant of Read Committed), we offer a simple probabilistic model that predicts the rate of integrity violations in our microbenchmark from configuration parameters. We validate the predictive model against measurements from the microbenchmark. The model identifies a region of the configuration space where a surprising inversion occurs: for these parameter settings, more integrity violations happen with Snapshot Isolation than with multi-version Read Committed, even though the latter is considered a lower isolation level.

Buddy System: Available, Consistent, Durable Web Service Transactions

“…To avoid snapshot isolation anomalies, we address blind writes and analyze the read log to determine if an anomaly could take place. Fekete et al [3] documented anomalies that must be avoided to turn a snap shot isolation schedule into a serialized schedule. We incorporate these requirements to support serializable execution.…”

Analyzing consistency properties for fun and profit

“…In contrast, the consistency properties we consider in this paper are the data-centric ones and are stronger and harder to verify. Fekete et al [11] investigate how often integrity violations are produced by varying degrees of isolation in database systems, but those violations are easy to verify.…”

How consistent is your cloud application?

“…In [15], the authors quantify violations of integrity constraints when a database system runs under read-committed or snapshotisolation. They analyze a microbenchmark that is composed of two related tables.…”

“…Fekete et al [15] provide quantitative information about the violations of integrity constraints when a centralized database system runs under various isolation levels. Wada et al investigated in [30] consistency properties for some cloud datastores, but they limited their study to a subset of client-centric consistency definitions such as monotonic-reads and read-your-writes.…”