Commutative Transformations in Multi-Model Databases

Recently there has been an increased interest in multi-model databases. This interest is motivated by big data technology and the need for horizontal scaling of stored data. To solve these problems, there have been developed toolkits that serve not only relational databases but also JSON, XML, spatial and key-value data structures. It is generally recognised by the users of such systems that the problem of multi-model data reconciliation still has no satisfactory solution. This paper proposes a heterogeneous data reconciliation method based on commutative transformations. Commutative transformations in databases were first investigated by L.A. Kalinichenko. The main distinction of this paper is the state bijectivity failure of different components of multi-model data. This is necessary since information is only partially duplicated in various system components.