BioCloud: A Framework for Efficient Biotechnology Applications Execution in Multiple Platforms of AWS Cloud
Coordinator: Alba Cristina Magalhães Alves de Melo
This research project aims to solve the problem of resource allocation and management in the Amazon AWS cloud for biotechnology HPC applications, minimizing execution time and maximizing availability, without violating SLA agreements, in order to run applications efficiently and cost-effectively. smaller financial costs for the user. In this sense, both permanent (on demand) and transient (spot) instances will be considered, as well as multiple types of computational resources (CPU, GPU and FPGA), increasing the range of alternatives explored.
The coordinated use of these multiple instances will be managed dynamically and autonomously. In this sense, the use of intelligent agents to assist in elasticity decisions (both horizontal and vertical) will be investigated. In terms of the public cloud computing model, in addition to the use of virtual machines only, the feasibility of using containers within instances will be studied in comparison with implementations using the container and Kubernetes services already offered by AWS itself and the concept of serverless computing with AWS Fargate versus AWS Lambda. However, the different approaches, technologies, and instances will be “hidden” from the end user through a scientific gateway (web portal) with a workflow manager harnessing the necessary cloud resources.
Team
Empresa Brasileira de Pesquisa Agropecuária (EMBRAPA) |
---|
Marcos Mota do Carmo Costa ![]() |
Natalia Florencio Martins ![]() ![]() |
Priscila Grynberg ![]() |
Roberto Coiti Togawa ![]() ![]() |