JetStream: Achieving Predictable Gossip Dissemination by Leveraging Social Network Principles

Gossip protocols provide probabilistic reliability and scalability, but their inherent randomness may lead to high variation in number of messages that are received at different nodes. This paper presents techniques that leverage simple social network principles enabling nodes to select gossip targets intelligently. The simple heuristics presented in the paper achieve a more uniform message overhead at each node, lowering the system-wide gossip traffic, while simultaneously reducing the latency of gossip spread (by up to 25%). We experimentally compare our system, called JetStream, against canonical gossip as well as gossip on the chord overlay. Intuitively, JetStream seeks to make gossip spread more deterministic and predictable, while still inheriting its scale and reliability. JetStream also provides an added benefit by reducing network bandwidth utilization with a low sustained rate of gossip injection