Towards low-latency I/O services for mixed workloads using ultra-low latency SSDs

Abstract

Low-latency I/O services are essential for latency-sensitive workloads when they co-run with throughput-oriented workloads in cloud data centers. Although advanced SSDs such as Intel Optane SSDs can offer ultra-low latency at the device layer, I/O interference among various workloads through the I/O stack can still significantly enlarge I/O latency. It is still an open problem to best utilize ultra-low latency SSDs in cloud computing environments.In this paper, we analyze the entire I/O stack and reveal that I/O interference is mainly attributed to resource contention in the SSD device, transactions commit in the file system, and costly process scheduling. To address these problems, we propose FastResponse, a holistic approach to use ultra-low latency SSDs for latency-sensitive workloads. First, we propose a new I/O scheduler at the block layer to throttle I/O requests of throughput-oriented workloads, and thus reduce the resource contention in the SSD device. Second, we develop a fine-grained journaling scheme to reduce the latency of transaction at the file system layer. Third, we redesign Completely Fair Scheduler (CFS) to promote the priority of latency-sensitive processes. We implement FastResponse in Linux kernel and evaluate it with several mixed workloads. Compared with the vanilla Linux and the state-of-the-art SelectISR, FastResponse can reduce the average response time of latency-sensitive workloads by 18-70% and 10-67%, respectively, and reduce the 99.9th percentile response time by 58-80% and 52-78%, respectively. Meanwhile, the performance degradation for throughput-oriented workloads is less than 6%. CCS CONCEPTS• General and reference → Performance; • Software and its engineering → Operating systems.

OctoFAS: A Two-Level Fair Scheduler That Increases Fairness in Network-Based Key-Value Storage

“…These approaches primarily address differences in memory access times incurred by interconnects between processors, e.g., NThread [40] and AASH [42] proposed thread migration as a means of avoiding contention in processor interconnects. There have been efforts directed toward load balancing in storage systems [44][45][46][47][48]. FastResponse [45] proposed several schemes applicable across the Linux storage stack to mitigate I/O interference between co-running low-latency I/O services.…”

“…There have been efforts directed toward load balancing in storage systems [44][45][46][47][48]. FastResponse [45] proposed several schemes applicable across the Linux storage stack to mitigate I/O interference between co-running low-latency I/O services. The blk-switch [44] achieved low latency and high bandwidth in the block storage layer by adopting a multi-queue design with load balancing and scheduling techniques.…”

Extending Memory Capacity in Modern Consumer Systems With Emerging Non-Volatile Memory: Experimental Analysis and Characterization Using the Intel Optane SSD

“…Previous works [100], [241], [260], [261], [262], [263], [264], [265] propose several techniques to mitigate block I/O latencies for fast NVMe devices. These techniques include software [100], [261], [262], [263], [264], [265] and hardware solutions [241], [260] to provide lower I/O access latency [100], [263], [264], page fault handling [260], and I/O scheduling [241], [261], [265].…”

“…Previous works [100], [241], [260], [261], [262], [263], [264], [265] propose several techniques to mitigate block I/O latencies for fast NVMe devices. These techniques include software [100], [261], [262], [263], [264], [265] and hardware solutions [241], [260] to provide lower I/O access latency [100], [263], [264], page fault handling [260], and I/O scheduling [241], [261], [265]. Even though these techniques are promising solutions to reduce the high block I/O latencies, they require substantial changes in the hardware and the software stack, which are outside the scope of this work, but can also be used in our proposed system.…”