Abstract High-concurrency systems often suffer from performance bottlenecks [1]. This is often caused by waiting and context switching caused by fierce competition between threads for locks. As a cloud computing company, we place great emphasis on maximizing performance. In this regard, we transform the lightweight spin-lock and propose a concise parameter fine-tuning strategy, which can break through the system performance bottleneck with minimal risk conditions. This strategy has been validated in Apache RocketMQ [2], a high-throughput message queue system. Through this strategy, we achieved performance improvements of up to 37.58% on X86 CPU and up to 32.82% on ARM CPU. Furthermore, we confirmed the method’s consistent effectiveness across different code versions and IO flush strategies, demonstrating its broad applicability in real-world settings. This work offers not only a practical tool for addressing performance issues in high-concurrency systems but also highlights the practical value of formal techniques in solving engineering problems.

As a team from Alibaba Cloud, we have developed and open-sourced RocketMQ, a cloud-native “messaging, eventing, streaming” real-time data processing platform that covers cloud-edge-device collaboration scenarios. During the development of RocketMQ, we also formulated RocketHA, a log-based storage high availability design theory that provides a robust solution for distributed log storage software used in industrial applications. RocketHA comprises six fundamental components that enable automatic cluster recovery from failures such as crashes and partitions. This design paradigm has been successfully implemented in the open-source RocketMQ. Our evaluation demonstrates that RocketHA ensures high availability, fast recovery, high throughput, and data loss prevention. We hope that RocketHA will inspire and guide the development of high-availability solutions for all log-based storage systems.