CIDER: Boosting Memory-Disaggregated Key-Value Stores with Pessimistic Synchronization

2026-04-03

AI Summary

This work addresses the performance bottleneck caused by excessive redundant I/O due to synchronization mechanisms in memory-disaggregated key-value stores under high concurrency, where the network becomes the limiting factor. The study introduces pessimistic synchronization into this architecture for the first time, proposing a contention-aware synchronization mechanism coupled with a global write coalescing strategy. This approach significantly reduces cross-node redundant I/O while preserving strong consistency. By adaptively balancing performance and correctness, the method achieves up to a 6.6× higher throughput compared to the state-of-the-art system on YCSB benchmarks, particularly excelling in low-contention scenarios.

Abstract

Memory-disaggregated key-value (KV) stores suffer from a severe performance bottleneck due to their I/O redundancy issues. A huge amount of redundant I/Os are generated when synchronizing concurrent data accesses, making the limited network between the compute and memory pools of DM a performance bottleneck. We identify the root cause for the redundant I/O lies in the mismatch between the optimistic synchronization of existing memory-disaggregated KV stores and the highly concurrent workloads on DM. In this paper, we propose to boost memory-disaggregated KV stores with pessimistic synchronization. We propose CIDER, a compute-side I/O optimization framework, to verify our idea. CIDER adopts a global write-combining technique to further reduce cross-node redundant I/Os. A contention-aware synchronization scheme is designed to improve the performance of pessimistic synchronization under low contention scenarios. Experimental results show that CIDER effectively improves the throughput of state-of-the-art memory-disaggregated KV stores by up to $6.6\times$ under the YCSB benchmark.

Problem

• memory-disaggregated
• key-value store
• I/O redundancy
• synchronization
• performance bottleneck

Innovation

• pessimistic synchronization
• memory-disaggregated KV store
• I/O redundancy
• write-combining
• contention-aware synchronization