SPRING: Improving the Throughput of Sharding Blockchain via Deep Reinforcement Learning Based State Placement
Keywords: blockchain, sharding, reinforcement learning, scalability
TL;DR: Using deep reinforcement learning to reduce the cross-shard transactions while maintaining the workload balance in the sharding blockchain, so that the overall throughput of the sharding blockchain can be improved.
Abstract: Sharding provides an opportunity to overcome the inherent scalability challenges of the blockchain. In a sharding blockchain, the state, and computation are partitioned into smaller groups, known as "shards," to facilitate parallel transaction processing and improve throughput. However, since the states are placed on different shards, cross-shard transactions are inevitable, which is detrimental to the performance of the sharding blockchain. Existing sharding solutions place states based on heuristic algorithms or redistribute states via graph-partitioning-based methods, which are either less effective or costly. In this paper, we present Spring, the first deep-reinforcement-learning(DRL)-based sharding framework for state placement. Spring formulates the state placement as a Markov Decision Process, which takes into consideration the cross-shard transaction ratio and workload balancing, and employs DRL to learn the effective state placement policy. Experimental results based on real Ethereum transaction data demonstrate the superiority of Spring compared to other state placement solutions. In particular, it decreases the cross-shard transaction ratio by up to 26.63% and boosts throughput by up to 36.03%, all without unduly sacrificing the workload balance among shards. Moreover, updating the training model and making decisions takes only 0.1s and 0.002s, respectively, which shows the overhead introduced by Spring is acceptable.
Track: Systems and Infrastructure for Web, Mobile, and WoT
Submission Guidelines Scope: Yes
Submission Guidelines Blind: Yes
Submission Guidelines Format: Yes
Submission Guidelines Limit: Yes
Submission Guidelines Authorship: Yes
Student Author: Yes
Submission Number: 423
Loading