Distributed Asynchronous Service Provisioning in Edge-Cloud Multi-Tier Networks

In edge-cloud multi-tier networks, datacenters must dynamically deploy and migrate services to meet diverse latency and computational constraints, especially for mobile users. Traditional solutions rely on centralized orchestration with full network visibility, which is impractical at scale and introduces reliability and privacy concerns. We propose DASDEC, a fully distributed and asynchronous framework for service placement and migration that operates without global coordination or private data exchange between providers. DASDEC enables local decision-making using lightweight control messages and supports cooperative placement across datacenters operated by distinct entities. Extensive simulations using real-world mobility traces show that DASDEC achieves performance very close to those obtained by optimal centralized solutions, while incurring negligible control-plane overhead (approx. 100 bytes per request). These results highlight DASDEC’s scalability, resilience, and practical applicability to federated edge-cloud systems.