Decentralized and Policy-Aware Serverless Orchestration for the Federated Web

Abstract

As Web services and Web-of-Things (WoT) applications increasingly span multiple administrative domains such as clouds, edges, and community-run platforms, centralized serverless orchestration frameworks struggle to handle heterogeneous policies, data governance constraints, and dynamic workloads. This situation prompts a new design space: federated and decentralized scheduling models for serverless computing that do not rely on a single global coordinator. In this paper, we explore an approach in which each administrative domain autonomously manages its own resources and complies with local regulations, while lightweight decentralized protocols enable these domains to cooperatively share load information, offer resource availability hints, and negotiate function placements. We illustrate how a small set of core mechanisms such as metadata exchange protocols, compliance-driven constraints, and trust-aware placement negotiations, could enable serverless functions to be flexibly placed and migrated across a federated ecosystem. Our experimental analysis suggests that decentralized coordination can maintain service-level objectives (SLOs) under diverse conditions, especially under load spikes, and align better with the policy and compliance requirements that characterize a heterogeneous Web environment. This paper charts a path forward for research on federated serverless orchestration, aiming to foster more flexible, scalable, and policy-aware infrastructure at global scale.