Automatic Minimization of Execution Costs of SPITS Programs in AWS

• Nicholas Okita UNICAMP
• Charles Rodamilans UNICAMP / UPM
• Tiago Coimbra UNICAMP
• Martin Tygel UNICAMP
• Edson Borin UNICAMP

Resumo

Cloud computing platforms offer a wide variety of computational resources with different performance specifications for different prices. In this work, we experiment how Spot instances and Availability Zones on the Amazon Web Services (AWS) could be utilized to reduce the processing cost. Not only that, but we propose an instance management algorithm in AWS to minimize the execution cost of programs implemented using the programming model Scalable Partially Idempotent Task System (SPITS). Our results show that the proposed method is able to identify and dynamically adjust the virtual machine types that offer the best cost per performance ratio.

Palavras-chave: Platinum, Computational modeling, Hardware, Web services, Task analysis, Google, Cloud computing, cloud, seismic processing, economics, aws, distributed computing