Abstract:
The article discusses the principles of data storage organization designed to maintain strict consistency between data replicas in distributed databases. Storing distributed data with strict consistency is necessary to solve the challenges of digitalization of the economy, whose subjects and decision- making centers are geographically separated. Typical representatives of the recently appeared NewSQL DBMS class are considered, and a conclusion is made about which data structures are used in them and why. The main problems of data organization are highlighted. The conclusion is made about the prospects of distributed databases with strict consistency. The direction of using key-value storages as a universal mechanism for storing distributed data is highlighted. Assumptions are made about why such repositories should be used and which mechanisms, for example, hashing, can help in organizing distributed storage on a large (up to hundreds of thousands) nodes of a distributed database data warehouse.
Keywords:organization of data storage, strict consistency, distributed databases, NewSQL, key-value stores, hashing.
Fulltext:
PDF file (265 kB)
First page:
PDF file