Andrea Di Prima
A Multi-Layered Framework for Reducing Execution Latency in Distributed Serverless Environments.
Rel. Alessio Sacco, Guido Marchetto, Vinay Sagar Prakash. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Informatica (Computer Engineering), 2026
Abstract
Serverless and edge computing have changed how online services are built. Serverless removes the always-on rented server, since developers write units of code that a platform starts on demand and stops when the work is done. Edge computing spreads that execution across data centres all over the world, so each request is handled a few milliseconds away from the user. Together they shorten the network path, but they also raise new challenges. Keeping a service fast now means shortening the distance between the user, the computation and the data at the same time, ensuring consistency and low execution latency. This thesis investigates where execution latency originates in Cloudflare RealtimeKit, a real-time media platform built around a WebRTC Selective Forwarding Unit and running on Cloudflare Workers and Durable Objects, and proposes a framework for reducing it.
The evidence comes from the running production system: percentile comparisons of REST and WebSocket response times, controlled reproduction of the observed failures, and measurement of every deployed change on real traffic
Relatori
Anno Accademico
Tipo di pubblicazione
Numero di pagine
Informazioni aggiuntive
Corso di laurea
Classe di laurea
Aziende collaboratrici
URI
![]() |
Modifica (riservato agli operatori) |
