Invariant Control in Eventually Consistent Databases

Due to the requirements imposed by data-intensive applications, NoSQL and NewSQL databases are becoming more present in the IT Market. These products provide alternative data models to the relational databases, and most of them are intrinsically distributed. These database management systems (DBMSs) relax consistency to favor availability and performance. However, applications that use NoSQL/NewSQL databases in distributed environments have to perform consistency control to avoid anomalies, such as inconsistent data and behavior. New approaches suggest the use of replicated data types (RDTs) to control conflicting updates. Another strategy is the use of different consistencies models for each type of operation, using first-order logic and theorem provers to supply the programmer with tools to classify consistency in operations while maintaining system invariants. Notwithstanding, the use of RDT or the descriptions of application integrity constraints in languages using first-order logic is still difficult to be used by programmers. Aiming to simplify the definition of most common database constraints, this paper proposes a mechanism to extract usual integrity constraints in an intermediate model, taking into account the semantics of invariances, using a mix of RDTs and first-order logic. The aim of this paper is to demonstrate how the proposed mechanism simplifies and guarantees safer programming with consistency control being performed at the application level.