Distributed Slicing in Dynamic Systems

Peer to peer (P2P) systems are moving from application specific architectures to a generic service oriented design philosophy. This raises interesting problems in connection with providing useful P2P middleware services capable of dealing with resource assignment and management in a large-scale, heterogeneous and unreliable environment. The slicing service, has been proposed to allow for an automatic partitioning of P2P networks into groups (slices) that represent a controllable amount of some resource and that are also relatively homogeneous with respect to that resource. In this paper we propose two gossip-based algorithms to solve the distributed slicing problem. The first algorithm speeds up an existing algorithm sorting a set of uniform random numbers. The second algorithm statistically approximates the rank of nodes in the ordering. The scalability, efficiency and resilience to dynamics of both algorithms rely on their gossip-based models. These algorithms are proved viable theoretically and experimentally.

💡 Research Summary

The paper tackles the problem of “distributed slicing” in large‑scale peer‑to‑peer (P2P) systems, a service‑oriented primitive that automatically partitions a network into groups (slices) each representing a controllable amount of a given resource (e.g., storage, bandwidth, CPU) while keeping the nodes inside a slice relatively homogeneous with respect to that resource. Traditional slicing approaches rely on centralized coordinators or heavyweight structural overlays, which do not scale well in heterogeneous, unreliable environments. To overcome these limitations the authors propose two gossip‑based algorithms that operate solely on local information, are fully asynchronous, and remain robust under high churn.

Algorithm 1 – Accelerated Random‑Number Sorting.
Each node draws a uniform random value in the interval