Optimised Latency in Cloud Systems: Multi-Level Caching through AWS-Based FDN

Challa, Harish (2025) Optimised Latency in Cloud Systems: Multi-Level Caching through AWS-Based FDN. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (849kB) | Preview

Abstract

Serverless computing has become a transformative model in cloud architecture by eliminating manual server management and enabling on-demand function execution. However, there are some challenges like cold start latency, repeated computation and stateless execution mostly degrade performance and hinder real-time responsiveness. This study proposes a Function Delivery Network (FDN) framework that extends the principles of Content Delivery Networks (CDNs) to function-level execution by using a multi-level caching mechanism to increase performance. The system uses a Level-1 in-memory Least Recently Used (LRU) cache for instant retrieval and a Level-2 persistent cache using AWS DynamoDB to store preprocessed type of results for minimizing redundant AWS Lambda invocations. This study has implemented through Flask, API Gateway and AWS Lambda within the AWS ecosystem. The FDN is experimentally validated using text and PDF files ranging from 40 KB to 2 MB. This study shows over 95% reduction in response latency, with Lambda executions averaging 492–2638 ms reduced to 0–21 ms through caching layers. These findings confirm that hierarchical caching enhances throughput, scalability, and cost efficiency while reducing cold starts.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Arun, Shreyas Setlur UNSPECIFIED
Uncontrolled Keywords:	Serverless Computing; Function Delivery Network; Multi-Level Caching; AWS Lambda; Latency Optimization
Subjects:	T Technology > T Technology (General) > Information Technology > Cloud computing Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Cloud Computing
Depositing User:	Ciara O'Brien
Date Deposited:	31 Aug 2026 12:27
Last Modified:	31 Aug 2026 12:27
URI:	https://norma.ncirl.ie/id/eprint/9692

Actions (login required)

View Item