Distributed Networking with cost effective solution

Hardik Rameshbhai Gondaliya

Bakalářská práce

Abstract:

This Study elucidates the challenges of broadcasting in networking, spotlighting the Broadcast Storm Problem (BSP) and its repercussions, particularly in Content Delivery Networks (CDN). Addressing redundancy, contention, and collision issues, the chapter examines broadcasting algorithms, such as simple flooding and probabilistic broadcasting, discussing their advantages and disadvantages. In the context of CDN, the text proposes modifications to these algorithms and introduces a clustered network structure with surrogates to alleviate BSP. Emphasizing the importance of reducing replication in CDN, the chapter underscores the potential efficacy of distance-aware and counter-based broadcasting approaches in mitigating BSP while enhancing reachability. The research focuses on optimizing Content Delivery Networks (CDN) to address flash crowd challenges by deploying replicas strategically using a population-based clustering algorithm. By modifying the K-Means algorithm based on population thresholds, the study aims to minimize deployment overhead and resource wastage. The algorithm redirects requests from underpopulated clusters to nearby surrogates, optimizing server utilization. Performance evaluation indicates comparable deployment costs with improved server efficiency in the population-based approach. This integration of distributed networking principles and CDN exemplifies a forward-looking strategy, promising cost-effectiveness, and performance excellence in the evolving digital landscape.

Keywords

Distributed Networking, Content Delivery Networks (CDN, Population-Based Clustering, -Means Algorithm Modification, Server Utilization Factor, Deployment Cost Optimization, Resource Efficiency, Request Distribution

Jazyk práce: angličtina

Datum vytvoření / odevzdání či podání práce: 2. 12. 2023

Obhajoba závěrečné práce

• Obhajoba proběhla 7. 2. 2024
• Vedoucí: Ing. Tomáš Vokoun
• Oponent: Jan Pavlík, Ph.D.

Česká zemědělská univerzita v Praze

Provozně ekonomická fakulta

Bakalářský studijní program:
Informatics