Design and Prototyping of Container-Enabled Cluster for High Performance Data Analytics

Abstract

With the rapid spread of cloud-native computing paradigm, the utilization of cloud has been maturing to accommodate versatile microservices-based application services with agility and scalability. However, in case of HPDA (high-performance data analytics) workload, the transition toward cloud-native-style computing infrastructure is still facing several hurdles to realize container-enabled clusters. To address technical constraints around bottlenecked inter-connections for overlay networking and storage access, in this paper, we design and prototype a container-enabled cluster that effectively and seamlessly integrate the hardware and software pieces of HPDA cluster.