EVALUATION OF NOSQL DBMS SCALABILITY FOR STORING DATA FOR PSYCHOLOGY WEB PLATFORM

The study examines the design and development of a web-based platform for conducting online psychological research. The authors substantiated the need for the development of a platform and described advantages over the pen-and-paper approach, both with respect to data collection and subsequent processing. The necessity of using NoSQL solutions for data storage is indicated, which is due to the large number of users of the system and the need to work with large volumes of data arriving at high speed. As the main advantage the possibility of sharding is considered, which is offered by NoSQL solutions by default. Replication is considered as an additional method for improvement of the platform availability and fault tolerance. An experiment was performed to analyze the scalability of NoSQL database management systems MongoDB and Cassandra, which estimated the speed of data recording with increasing the number of nodes (from 1 to 3 nodes). The performance gain is noticeably better with an increase in the write volume to about 10K per query for MongoDB and 100K per query for Cassandra. In view of Cassandra's high requirements for the qualifications of engineers, MongoDB was chosen as a solution for developing a web-based platform for conducting psychological research. (C) 2017 Published by Future Academy www.FutureAcademy.org.UK