Traffic Redundancy in Blockchain Systems: The Impact of Logical and Physical Network Structures
Yinghao Zhang, Xiao Fan Liu · 2021
The current designs of Internet-based blockchain systems generate huge traffic redundancy, thereby reducing the efficacy of the system. This paper simulates the Bitcoin blockchain network traffic in a realistic setting and studies the impacts of both logical and physical network structure on the traffic redundancy. Firstly, we construct an Internet AS layer topology graph and map the Bitcoin nodes to the ASes. Then, Bitcoin 's neighbor discovery algorithm is used to connect Bitcoin nodes together logically. Finally, we simulate the Internet traffic generated by Bitcoin 's flood-based broadcast algorithm in the AS layer topology graph. Specifically, the traffic that a message generates is calculated by the number of ASes a message passes through times the message size. The difference between the total and effective traffic (i.e., the theoretical minimum) is considered traffic redundancy. Simulations show that traffic redundancy is positively correlated with the Bitcoin network scale and the number of peer neighbors for each node. Meanwhile, the underlying physical network topology, especially the average routing path length between the Bitcoin nodes, determines the traffic redundancy as well. Our findings imply that one can optimize the blockchain network's throughput by altering the underlying physical network, e.g., migrating from the Internet to a satellite environment.