The Research On The Eventual Consistency Method Based On The Vector Clock For NoSQL

In digital era, big data provides more commercial value and developmentopportunities for enterprises, but also put forward higher requirements to manage andanalyze the data effectively: huge amount of data throughput, lower access delay, strongavailability, scalability and less operating costs etc. The non-relational database that usesthe eventual consistency technology well meets these requirements. By weakeningconsistency, high scalability and high availability is achieved. Hence, NoSQL meets thechallenges of big data better.Eventual consistency is an important indicator to measure the performance of NoSQL.It plays an important role in the development and application of NoSQL. In all solutions toeventual consistency of NoSQL, vector clock technology is used most widespread. Theeventual consistency model based on vector clock can distinguish data’s time sequenceeffectively. To some extent, it ensures the ordering of operations and the consistency ofthe data.By studying algorithm rules of the eventual consistency model based on vector clockfurther, the model in solving eventual consistency of NoSQL still existing some problems:information used to control the consistency has a linear growth with the number of nodes.The large number of nodes would lead to large amount of message transmission so that thenetwork may be over-loaded. It can’t distinguish the timing relationship whenencountering update operations with non-causal relationship, which may lead to datacovered by mistake, lower read performance and higher error rate, etc. To solve theseproblems, the eventual consistency model based on global vector clock incrementalsynchronization is putted forward, and introduces in detail its basic idea and algorithm.Finally, the model is proved to be correct by theory. And through examining severalexamples, it is proved to decrease greatly the amount of information that used to controlthe consistency. It can also solve the problem of guaranteeing order of operations withnon-causal relationship. Hence, it is a feasible and effective solution.