An Improved Gossip Algorithm Based on Semi-Distributed Blockchain Network

Xiaowei He, Yiju Cui, Yunchao Jiang

2019 · 26 citations · 10 references

Concepts

TL;DR

Blockchain projects increasingly adopt semi‑distributed P2P networks, yet the standard gossip algorithm suffers from fixed target‑node selection that causes duplicate messages, redundancy, and inefficient data synchronization. This paper proposes an improved HNA‑Gossip algorithm to mitigate duplicate node selection during data synchronization in semi‑distributed blockchain networks. The algorithm dynamically records historical node information to adjust target‑node selection probabilities and reduce redundancy. Simulations demonstrate that HNA‑Gossip outperforms the original gossip algorithm across multiple performance metrics.

Abstract

With the continuous development of blockchain technology, more and more blockchain projects use semi-distributed P2P network structures. Although original gossip algorithm can be devoted to data synchronization in semi-distributed blockchain network, it can not be well applied to actual network environment. Since the probability of selecting a target node during data synchronization is fixed, it is inevitable that a message can be sent to a duplicate node. It will not only cause a lot of redundant messages, but also bring inefficient data synchronization. To address this problem, this paper proposes an improved HNA-Gossip algorithm which can reduce the probability of selecting duplicate nodes to send messages by recording historical node information dynamically. The simulation results show that, compared with the original gossip algorithm, various aspects of HNA-Gossip algorithm perform better.

References

10