LOAD BALANCING CONSISTENCY IN A DISTRIBUTED DATASTORE

Abstract

The subject of the article’s research is the CAP-guarantees of distributed datastore, particularly availability and consistency. The goal is to design an approach that will become an instrument to balance consistency CAP-guarantee for any business needs still maintaining appropriate availability guarantee. The algorithm could be integrated to datastore infrastructure as one of distributed datastore components that must stand on top of or integrated in database middleware standing on the path to node database instance and actual query execution. To achieve that, the following problems were solved in the paper: the simulation models for approaches have been implemented, actual possibility to implement an algorithm following specific approach has been investigated. The following methods were used to implement such solutions: UML modeling, computer model implementing the simulation of the designed algorithms, carrying experiments on the implemented models. Carried out experiments resulted in capability to estimate the complexity and possible performance and make conclusions choosing one of optimal approaches to be designed further. As a conclusion, the optimal designed and estimated approach of balancing consistency and availability is ready and it was the purpose of this paper. It could be applied as one of basic components on the design distributed datastores stage, so that balanced guarantees of distributed system reliability could be achieved at the earlier stage of business needs implementation