ThespisTRX : initial results for causally-consistent read transactions

Abstract

Data Consistency defines how usable a data set is, and different levels of data consistency have been proposed. Causal consistency is one of these. It is the strongest type of consistency that can be achieved when data is stored in multiple locations, and fault tolerance is desired. Thespis is a middleware that innovatively leverages the Actor model to implement causal consistency over a DBMS, whilst abstracting complexities for application developers behind a REST interface. Following the evaluation of correctness, performance and scalability of Thespis, in this paper the authors illustrate how a business application can be guaranteed causal consistency, but still encounter Time-To-Check-Time-To-Use (TOCTOU) race conditions. The authors subsequently present the design of ThespisTRX, which builds upon, and extends, the Thespis middleware to offer read-only transaction capabilities, allowing clients to read a causally-consistent version of multiple data entities. A correctness analysis illustrates how the design avoids TOCTOU race conditions, and empirical tests show the efficacy of the approach.