Automated tools for the development of microservice compositions for hybrid scientific computations

Abstract

In recent years, a significant amount of research is focused on the development of tools for creating composite web-services for solving both business and scientific complex problems. This study discusses tools for building compositions or ensembles of microservices (depending on the method of integration) developed based on the HPCSOMAS framework. These tools are oriented on the application in a package of applied microservices for solving computationally complex problems of structural analysis and parametric synthesis of controlled dynamic systems in a heterogeneous high-performance computing environment. In particular, binary dynamic systems are studied using the Boolean constraint method for both their qualitative analysis and synthesis of laws to control these systems. Creating and executing composite services is carried out on a semantic peer-to-peer network of agents. The HPCSOMAS framework supports two modes of these processes, both the static creation and application of a composite service based on the procedural formulation of the problem and dynamic, based on the declarative formulation. In the first case, agents deployed on the network perform hierarchical control over the execution of the composition of microservices, in the second case, decentralized asynchronous management of the ensemble of microservices. Both operating modes are automated, and the validity of the resulting composite service is checked based on a logical approach. The tools are aimed both at a professional programmer and the end-user, a specialist in the subject domain. The HPCSOMAS framework supports the execution of composite microservices in a hybrid computing infrastructure, which includes both cloud and on-premises resources.