ThespisTRX

• DOI: 10.4018/ijitwe.2020010101
• Publication Date: 2019-11-04T13:29:02Z
• Authors: Carl Camilleri; Joseph G. Vella; Vitezslav Nezval
• Keywords: Causal consistency; Cache coherence; USable; Weak consistency; Eventual consistency

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