Latency reduction techniques in real-time processing pipelines with edge computing and serverless

Abstract

This article examines latency reduction techniques in real-time processing pipelines, with an emphasis on the convergence of Edge Computing and Serverless architectures. The discussion adopts a historical-evolutionary approach, tracing the path from the consolidation of cloud computing—defined by NIST as on-demand access to shared, elastic, and measurable computational resources—to the decentralization promoted by edge architectures. It is argued that latency reduction depends not only on greater computational capacity but also on architectural decisions about processing placement, functional granularity, asynchronous communication, data partitioning, observability, and cold-start mitigation in serverless platforms.