Implementation of Serverless Architecture on Public Cloud: A Hybrid Approach to Optimizing Cost and Performance for Data-Intensive Workloads

Thomas, Alfi (2025) Implementation of Serverless Architecture on Public Cloud: A Hybrid Approach to Optimizing Cost and Performance for Data-Intensive Workloads. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (769kB) | Preview

Abstract

With the increase in demand due to the rise of data-rich applications have increased demands for cloud architecture that are both scaleable and cost effective. One of the major forms of cloud services is serverless computing, which provides an easy way to manage data with a pay-per-use billing model. But because of the limitations on execution time, memory and processing capabilities, serverless computing is not suited for large-scale data compute workloads. In this paper, build a new hybrid serverless model based on Amazon Web Services (AWS) public cloud as a means to solve this problem. By analyzing the data file sizes, determine which compute engine to use based on thresholds: for smaller tasks, use AWS Lambda and for larger resource intensive tasks, use AWS Fargate. This hybrid model will allow for a unified processing pipeline that utilizes the cost-effective characteristics of serverless functions for the majority of tasks but provides scalability into a containerized environment when serverless functions cannot handle the excessive loads. The system was developed using Infrastructure as Code (IaC) methodology in Python with AWS Boto3 SDK for the code development and all test scenarios were conducted to verify success through small, medium and large data files. In our evaluation, found that the hybrid architecture was successful in automating the decision-making process for the compute engine, maintaining a 100% success rate across all tests, and generating total operational costs estimated to be only a small percentage of a cent. This research shows that the hybrid serverless model is a very effective means to support the management of heterogeneous workloads for large amounts of data in a public cloud environment.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Samarawickrama, Yasantha UNSPECIFIED
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Cloud computing
Divisions:	School of Computing > Master of Science in Cloud Computing
Depositing User:	Ciara O'Brien
Date Deposited:	01 Sep 2026 11:49
Last Modified:	01 Sep 2026 11:49
URI:	https://norma.ncirl.ie/id/eprint/9744

Actions (login required)

View Item