Serverless Workflows for Containerised Applications in the Cloud Continuum

Sebastián Risco,Germán Moltó,Diana M Naranjo,Ignacio Blanquer

doi:10.1007/s10723-021-09570-2

Abstract

This paper introduces an open-source platform to support serverless computing for scientific data-processing workflow-based applications across the Cloud continuum (i.e. simultaneously involving both on-premises and public Cloud platforms to process data captured at the edge). This is achieved via dynamic resource provisioning for FaaS platforms compatible with scale-to-zero approaches that minimise resource usage and cost for dynamic workloads with different elasticity requirements. The platform combines the usage of dynamically deployed auto-scaled Kubernetes clusters on on-premises Clouds and automated Cloud bursting into AWS Lambda to achieve higher levels of elasticity. A use case in public health for smart cities is used to assess the platform, in charge of detecting people not wearing face masks from captured videos. Faces are blurred for enhanced anonymity in the on-premises Cloud and detection via Deep Learning models is performed in AWS Lambda for this data-driven containerised workflow. The results indicate that hybrid workflows across the Cloud continuum can efficiently perform local data processing for enhanced regulations compliance and perform Cloud bursting for increased levels of elasticity.

Highlights

Cloud computing has become in the last decade the premier option for virtualised computing
In order to prove the effectiveness of hybrid serverless workflows for processing data produced at the edge, the times obtained after 5 workflow runs to process a single video were measured
This increment is caused by the uploading of the images to the input bucket of the second function, which in the first case was located in the same cluster, while in the hybrid workflow is on Amazon S3, so the files must be uploaded via the Internet

Summary

Introduction

Cloud computing has become in the last decade the premier option for virtualised computing It has increased hardware resource utilization and provided the ability to execute disparate computing workloads with complex requirements on shared computing infrastructures. Initial service delivery models, such as Infrastructure as a Service (IaaS), were exemplified by public Cloud services such as Amazon EC2 [4] and on-premises Cloud Management Platforms (CMPs) such as OpenStack [48]. These were later extended to accommodate additional models such as Platform as a Service (PaaS) and, more recently, Functions as a Service (FaaS). FaaS aims to rise the level of abstraction for application developers at the expense of relying on the infrastructure provider for automated elasticity, efficient virtual infrastructure provisioning and improved resource allocation.

Objectives

Results

Conclusion