Analysis of the Logical Consistency in Cassandra

Abstract

Most methods to test the consistency of data in relational databases are based on checking the integrity of the data through testing of the schema. However, in NoSQL databases, there is no schema and the data is denormalized, so we need new methods to test the consistency of the database. One of these NoSQL databases is Cassandra, where each table is created to satisfy one query. As the same information could be retrieved by several queries, this information may be found in several tables. In these cases, whenever there is a change of the values of a row in a table, the consistency could be affected. In this thesis, we propose a preventive approach to this problem, where we test the consistency of the data by doing a static analysis of the Cassandra tables when there is a change of the data that could affect the consistency. To achieve this, we will use a conceptual model to help us in the static analysis.