Sharding is a promising solution to enhance blockchain scalability. While deploying more shards of smaller sizes for a given network scale can significantly boost performance, it also heightens the risk of shard failures. In many existing sharding systems, the failure of a single shard can compromise the entire system. Therefore, to ensure safety, current systems often require each shard to contain hundreds of consensus nodes to prevent crashes, adversely affecting scalability. In this paper, we propose Levee, a blockchain sharding system capable of tolerating shard failures. When a shard malfunctions, Levee can swiftly detect, isolate, and autonomously recover the faulty shard, allowing other shards to operate without interruption. This fault-tolerance feature enables Levee to reduce shard sizes by 73.7% and increase the number of shards by 3.25 times, all without sacrificing security. When tested in a scenario with 7000 nodes, Levee demonstrated a 14.34 times increase in throughput and a 65% decrease in transaction latency compared to traditional non-fault-tolerant sharding systems.

Summary Multivariate time series classification holds significant importance in fields such as healthcare, energy management, and industrial manufacturing. Existing research focuses on capturing temporal changes or calculating time similarities to accomplish classification tasks. However, as the state of the system changes, capturing spatial‐temporal consistency within multivariate time series is key to the ability of the model to classify accurately. This paper proposes the MSTformer model, specifically designed for multivariate time series classification tasks. Based on the Transformer architecture, this model uniquely focuses on multiscale information across both time and feature dimensions. The encoder, through a designed learnable multiscale attention mechanism, divides data into sequences of varying temporal scales to learn multiscale temporal features. The decoder, which receives the spatial view of the data, utilizes a dynamic scale attention mechanism to learn spatial‐temporal consistency in a one‐dimensional space. In addition, this paper proposes an adaptive aggregation mechanism to synchronize and combine the outputs of the encoder and decoder. It also introduces a multiscale 2D separable convolution designed to learn spatial‐temporal consistency in two‐dimensional space, enhancing the ability of the model to learn spatial‐temporal consistency representation. Extensive experiments were conducted on 30 datasets, where the MSTformer outperformed other models with an average accuracy rate of 85.6%. Ablation studies further demonstrate the reliability and stability of MSTformer.

Sharding is one of the most promising technologies for significantly increasing blockchain transaction throughput. However, as the number of shards increases, the ratio of cross-shard transactions in existing blockchain sharding protocols gradually approaches 100%. Since cross-shard transactions consume many times more resources than intra-shard transactions, the processing overhead of cross-shard transactions already accounts for the majority of the total overhead of the sharding system. There is a very large gap between the transaction throughput of the sharding system and its theoretical upper limit. In this article, we propose Estuary, a novel low cross-shard blockchain sharding protocol. Taking the state model as an entry point, Estuary designs a multi-level state model and state splitting and aggregation mechanism. It decouples the identity and quantity of state units, enabling transactions between users to be completed within one shard. Only when the state quantity for all shards of a user is insufficient a small number of cross-shard transactions are required. On this basis, we propose a community overlap propagation algorithm for sharding. It defines the users’ belonging coefficients of each shard and optimizes the state distribution so that the state distribution can better match the transaction characteristics between users. Finally, we develop an analysis framework for the sharding protocol and experiment with real Bitcoin transactions. The evaluation results show that compared to the state-of-the-art sharding protocol, Estuary reduces the ratio of cross-shard transactions by 88.54% and achieves more than 1.85 times the throughput improvement (92.98% of the theoretical upper limit).

Broadcasting plays a vital role in the consensus mechanisms of blockchain systems, since the consensus of each block must wait until the previous block is received by (nearly) all the nodes in the blockchain systems. Therefore, optimizing the performance of broadcasting can significantly improve the performance of the blockchain system. However, compared with other traditional P2P applications such as file downloading or video delivery, the broadcasting in blockchain has two new requirements, namely low redundancy and low propagation latency, which all the existing mechanisms (e.g. flooding, structural DHT etc.) can not meet well. In this paper, we propose Swift, a new broadcasting mechanism for blockchain systems. It optimizes the P2P topology construction and broadcast algorithm in the structured network based on unsupervised learning and greedy algorithm, effectively reducing the propagation latency of the blockchain P2P network while avoiding the waste of redundant bandwidth. We implemented a prototype of Swift and evaluated its performance on a testbed network that consists of 1000 blockchain nodes. The experimental findings show that Swift can reduce propagation latency by 19.8% with similar bandwidth consumption, generating an 18% increase in the throughput performance of the blockchain. Finally, with the increase in connections, Swift can simultaneously achieve low latency and maintain a relatively stable redundant bandwidth waste, instead of linearly increasing in flooding.