FairShare
Project Profile

FairShare

A decentralized resource sharing protocol for edge devices. Utilizes federated learning to allow edge clusters to cooperatively manage power limits without central cloud orchestrators.

Edge Computing Federated Learning
View Repository

Dynamic Edge Orchestration

FairShare is a decentralized, cooperative resource sharing protocol designed to enable edge compute nodes to optimize cluster power distribution and resource load-balancing autonomously. By avoiding centralized cloud orchestrators, FairShare eliminates single points of failure and significantly minimizes network overhead.

Core Technologies

  • Federated Resource Allocation: Nodes communicate model parameters regarding their power margins and compute usage rather than raw traffic data.
  • Adaptive Power Control: Keeps total edge cluster power usage below set bounds while ensuring critical low-latency edge jobs receive optimal computing resource shares.
← Back to all projects