DATA CONSISTENCY MODELS IN APPLICATIONS WITH MICROSERVICE ARCHITECTURE

Abstract

Individual database transactions can easily meet ACID requirements and ensure strong consistency, but distributed transactions have a number of limitations to address. Hard transactions are are found to be fully ACID compliant, while using flexible transactions isolation is not fully guaranteed. In practice, the isolation requirement is waived to some extent to ensure high throughput and system performance. Flexible transactions generally adhere to the Basic Availability, Flexible State, Eventually Consistency (BASE) theory. The BASE theorem is an extension of the CAP theorem. This is the balance between consistency and availability in CAP. According to the BASE theory, we cannot achieve strong consistency, however, each program can achieve ultimate consistency by using an appropriate method according to its own characteristics. CAP shows that a distributed system cannot achieve consistency, availability, and distribution resilience at the same time. It is worth to pay attention at the stage of system design. It was found that hard transactions are strive for persistent consistency, and so they sacrifice high availability. Flexible transactions sacrifice consistency in exchange for high system availability.