CXDVirt: Low Latency Kernel Module Based CXL-SSD Emulation

2026-10-06

AI Summary

This study addresses the latency distortion in existing virtual machine-based CXL-SSD simulators caused by architectural overhead, which impedes accurate hardware performance evaluation. We propose a lightweight CXL-SSD emulation framework implemented as a Linux kernel module. By entirely eliminating the virtual machine layer to enable native host access, the approach leverages the devdax interface and a custom page fault handling path to directly inject NAND latencies. Experimental results demonstrate that this method reduces miss latency by 1.4–1.6× and improves P99.9 tail latency by up to 4.1×. In the absence of NAND traffic, performance closely approximates that of remote DRAM. This work provides a high-fidelity evaluation platform for advancing research on CXL storage strategies.

Abstract

CXL-SSDs promise memory-semantic access to NAND-scale capacity through a small in-device DRAM cache, but scarce hardware makes emulation central to CXL-SSD research. Cylon, a state-of-the-art emulator, runs workloads in a VM, intercepting DRAM misses to inject NAND latency. We show this VM machinery adds over 3$μs$ per miss, exceeding high-performance NAND read latency, while repeated VM exits inflate tail miss latencies several-fold. We present CXDVirt, a kernel-module-based CXL-SSD emulator that removes the VM entirely: the device is exposed through host devdax, hits run natively via the MMU, and a custom page-fault path models NAND latency on misses. CXDVirt preserves CXL-SSDs' bimodal latency profile while reducing per-miss latency 1.4-1.6$\times$ (4.1$\times$ at P99.9); with eight threads its miss latency only doubles, against 8-9$\times$ for Cylon behind QEMU's global lock. CXDVirt stays within 5% of remote DRAM without NAND traffic, where Cylon is 3.1-3.6$\times$ slower, and enables configurable eviction and prefetch policy studies.

Problem

• CXL-SSD emulation
• low latency
• tail latency
• virtual machine overhead

Innovation

• CXL-SSD Emulation
• Kernel Module
• Page-Fault Path
• Low Latency
• devdax