THE PLATFORM FOR CREATION OF EVENT-DRIVEN APPLICATIONS BASED ON WOLFRAM MATHEMATICA AND APACHE KAFKA

• Published in: Sučasnij stan naukovih doslìdženʹ ta tehnologìj v promislovostì (Online) no. 2 (16); pp. 12 - 18
• Main Author: Zolotariov, Denis
• Format: Journal Article
• Language: English
• Published: Kharkiv National University of Radio Electronics 06.07.2021
• Subjects: cloud technologies; event-driven applications; Kafka; mathematical processor; queue manager; saving resources and funds
• ISSN: 2522-9818; 2524-2296
• DOI: 10.30837/ITSSI.2021.16.012

The article is devoted to the study and development of the mechanism of interaction between Wolfram Mathematica programs and Apache Kafka queue to provide the ability to build event-driven applications based on it. The subject of the research is the practical principles of building a mechanism for interaction between Wolfram Mathematica and Apache Kafka through a proxy-server. The purpose of the article is to develop and substantiate practical recommendations regarding the formation of proxy-server and a mechanism for its work to publishing messages to the Apache Kafka queue and reading messages from it for programs of the mathematical processor Wolfram Mathematica, which will make it possible to build event-driven applications. The tasks are: to determine the mechanism of such interaction, prove the choice of tools for its implementation, create and test the obtained results. The research used the following tools: Apache Kafka, Kafkacat, servers Ubuntu 20 LTS, the method of developing the Wolfram Mathematica package. The results of the research: the mechanism of interaction between Wolfram Mathematica and Apache Kafka through a proxy-server was determined and the corresponding toolkit was created on its basis in the form of two Mathematica packages, which are built on using bash-scripts, Apache Kafka and third-party Kafkacat software. The first - for use on the end user's computer, the second – on a compute server with a remote Mathematica kernel. It is confirmed that the Mathematica processor is currently not suitable in its pure form for real-time data analysis. Conclusions. Practical recommendations have been developed and substantiated regarding the formation of the mechanism of interaction between the Wolfram Mathematica mathematical processor and the Apache Kafka queue manager through a proxy-server for the possibility of working in two directions with the queue: publishing messages and reading them. A toolkit for such interaction in the form of Mathematica packages has been created, their capabilities have been demonstrated. The economic benefit of using the described tools is shown. Future ways of its improvement are given.