Collaborative, distributed and scalable platform based on mobile, cloud, micro services and containers for intensive computing tasks

Abstract

Compute-heavy workloads are traditionally run on x86-based HPC platforms and Intel, AMD or Nvidia GPUs; these require a high initial capital expense and ongoing maintenance costs. ARM-based mobile devicesoffer a rad-ically different paradigm with substantially lower capital and maintenance costs and higher gains in performance and efficiency in recent years. When compared to their x-86 brethren, they have become ubiquitous in consumer markets and are making steady gains in the server market. Given this shifting computer paradigm, it is conceivable that a cost- and power-efficient solution for our world’s data processing would include those very same ARM-based mobile devices while they are idling. Given that context, we developed and deployed an auto-scalable, distributed and redundant platform on the basis of a cloud-based service managed via container orchestration and microservices that are in charge of recycling and optimizing these idle resources. We tested the platform performing distributed video compression. We concluded the system allows for improvements in terms of scalability, flexibility, stability, efficiency, and cost for compute-heavy work-loads.