RESEARCH OF METHODS OF SOFTWARE IMPLEMENTATION OF THE COSMOS DB API ON THE .NET PLATFORM

Abstract

Large number of developers use the .NET platform to create applications that work with databases today. In turn, Cosmos DB is becoming an increasingly popular choice as a NoSQL storage for such databases. Cosmos DB is a flexible and scalable system, and the correct selection of the appropriate API during software implementation can significantly affect the performance of the programs themselves. Cosmos DB provides different APIs for working with different types of databases, such as SQL databases, running MongoDB or Cassandra. In turn, each of these APIs can be used using various methods of software implementation. The subject of research is software implementations on the .NET platform for various Cosmos DB APIs. When choosing the most suitable Cosmos DB API on the .NET platform, developers can be helped not only by the documentation, but also by the results of experimental studies of these APIs, which in turn will improve the quality of the code and the performance of the systems themselves. The goal of the work is to increase the efficiency of software development on the .NET platform. That use the Cosmos DB API, by developing a recommendation for the selection of software implementation methods for these APIs based on the results of their experimental research. The task: to investigate and compare the methods of software implementation of the Cosmos DB API through an experimental study of the performance of different types of queries on these software solutions; analyze the obtained results and develop recommendations for the use of methods. Methods: multi-criteria analysis of Cosmos DB API, logical data modeling, experimental research. Results: developed software solutions based on the use of CosmosClient, Entity Framework Core for Cosmos DB API for NoSQL and based on MongoClient for Cosmos DB API for MongoDB. A series of experiments and measurements of performance metrics for each of the software solutions were conducted, the obtained results were analyzed, and recommendations were offered for using the considered methods of software implementations of the Cosmos DB API on the .NET platform. Conclusion: In general, the choice of software approach depends on the specific task, but experiments have shown that CosmosDB API for NoSQL using CosmosClient is the best choice for small projects, and using the Entity Framework Core Cosmos is suitable for more complex projects with larger volumes of data and complex queries. If MongoDB is used in the project, then the corresponding solution using MongoClient is a better option than Cosmos DB API for NoSQL.