Investigating the new concept of building a dynamic problem-oriented software architecture in Grid-Cloud environment with elements postbinary computing

SOC обчислювальна платформа

Topic: Methodology development of problem-oriented software architecture building by combining web-based, grid and cloud services.

A new methodology of distributed problem-oriented software architecture is developed for solving scientific and engineering problems of high complexity. Unlike existing approaches, it allows you to dynamically synthesize application software from available network software modules via Web-services interfaces (both for SOAP-services and REST-services) and utilizing high-performance computing resources both grid network and cloud infrastructures to meet specific tasks and requirements of the end user, who is not a specialist in information technology.

Tools for integrating different types of resources in high-performance software architecture were developed in the form of "gateway services" perform for calculations that hide the low-level details of interaction with a particular cloud and grid infrastructure.

The decentralized model of applied and generic microservices-agents "ecosystem" is proposed which forms the basis of dynamic distributed architecture with low components connectivity. Microservices -agents interact with each other to solve the problems through the exchange of standardized messages based on the knowledge base of their "ecosystem", creating dynamic workflows on user request.

A repository of computational services for complex systems modelling is built. The method for setting performance computing tasks using semantic web technology is developed and implemented, which, in contrast to existing approaches, allows to formulate the ultimate goal of computing and to refine ways to achieve it for an user who does not have the skills programmer. Specific workflow of Web services that will be involved in a particular computing scenario realization is built automatically and is able to dynamically adjust according to changes in microservices ecosystem. The research results are implemented in practical development of the complex computing applications.

AttachmentSize
PDF icon 2016_2710.pdf746.27 KB