Optimizing Cold Start Latency in Serverless Architectures through Edge-based Resource Allocation

Ademola, Oluwaseyi Ezekiel (2024) Optimizing Cold Start Latency in Serverless Architectures through Edge-based Resource Allocation. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (4MB) | Preview

Abstract

Serverless computing architecture enables the development and deployment of applications without the management or knowledge of the underlying infrastructure allowing developers to focus on building business logic while delegating resource management operations to the cloud service providers (CSPs). This simplification means that resources are allocated on demand and charges are on a pay-as-you-use basis making it an affordable option for most users. CSPs achieved this execution model by leveraging containerization whereby application codes are deployed in an isolated, stateless environment that provides the necessary resources for execution. Upon completion, these resources are released, requiring the fresh provisioning of resources for subsequent executions. The initialization of a new container, however, incurs a delay referred to as cold start latency, hence limiting the applicability of this model in latency-sensitive applications. This research addressed the problem using a location-aware algorithm that harnessed user mobility patterns and geofence-based prewarming to schedule containers and execute functions closer to the user. The setup was executed through simulation using real-world serverless functions and mobility datasets provided by Microsoft, and the result recorded a 70% reduction in cold start latency and a 76% decrease in resource consumption of function execution during peak and non-peak periods.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Hamza Ibrahim, Ahmed 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 Q Science > QA Mathematics > Algebra > Algorithms > Computer algorithms
Divisions:	School of Computing > Master of Science in Cloud Computing
Depositing User:	Ciara O'Brien
Date Deposited:	14 Jul 2025 13:47
Last Modified:	14 Jul 2025 13:47
URI:	https://norma.ncirl.ie/id/eprint/8076

Actions (login required)

View Item