Abstract

This paper shows that performance of the learning management systems heavily depends on the choice, made during designing, of architectural solution for storage and processing of data. Based on analysis of evolution of the various architectural solutions dur-ing the information system design, beginning with monolith platform and ending with decentralized microservices, it has been de-termined that architecture based on microservices for a server side with code-level isolation and database-level decentralization for components is proved to be effective solution for high-performance system complexes for learning management system. However, for implementation of polyglot persistence concept based on multiple database managementsystems with various logical schemas, there is also a need for developing an information technology to support such architectural solutions. It has been shown that the de-velopment of databases for such learning management system, that operate with a large amount of various information, consists of the stages of conceptual, logical and physical modeling, and, precisely during the creation of logical models the requirements for the storage and processing of data, that are used by the selected entities for the implementation of business functions, are determined. The peculiar properties of using relational and non-relational database management systems such as: document, key-value, graph and column storages have been examined and analyzed in detail. A method for automated selection of logical data models based on initial information about a limited context has been developed, then used to develop a classifier. The efficiency of the classifier was tested on a dataset for two hundred thirty entities. As a result of the experiment, the reliability of the classification was ninety-three percent. The advantages of the developed information technology are shown on the example of designing JustStart learning management system. Analysis of the stress testing results of the developed system shows that due to the distribution of the load between the three databases, its average response time with simultaneous operation of one hundred fifty users was one point two seconds. At thesame time, simulation of the same number of users with only one database management system, the response time increased and the aver-age was approximately two point six seconds. Thus, the use of the developed information technology of supporting architectural solutions for organizing storage of large volumes of diverse data according to the polyglot persistence concept, that allowed to design and implement learning management system, the performance of which, if it is used simultaneously by a large audience, is on aver-age twice as fast as the average educational resource on the market.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call