A Comparative Analysis of Performance and Usability on Serverless and Server-Based Google Cloud Services

Abstract

The purpose of the study is to compare and analyze the performance and usability of a containerized mobile-web application on serverless and server-based services in Google Cloud. The application allows users to view the real-time location of San Antonio VIA buses on specific bus routes and obtain estimated arrival times for chosen bus stops. The application was hosted in Google Cloud’s serverless Cloud Run and App Engine as well as server-based Compute Engine. In analyzing the performance, Google Cloud’s Monitoring services were used to measure performance metrics, including latency, throughput, and CPU utilization. The documentation and learning curve of the services were also evaluated to access their usability. The study’s main contribution is to enhance our understanding of the performance and usability of utilizing Google’s serverless and server-based services for hosting containerized applications. It aims to help individuals in selecting the most suitable cloud service for their hosting requirements. The results of the study indicated that Cloud Run outperformed App Engine and Compute Engine in terms of both performance and usability. Cloud Run demonstrated the lowest latency of 47.24 ms and highest throughput with sent and received rates of 9.47 k/s and 11.15 k/s, respectively. Additionally, it exhibited the lowest container startup latency of 475.22 ms and CPU utilization of 1%.