Tiny-Tail Flash

Abstract

Flash storage has become the mainstream destination for storage users. However, SSDs do not always deliver the performance that users expect. The core culprit of flash performance instability is the well-known garbage collection (GC) process, which causes long delays as the SSD cannot serve (blocks) incoming I/Os, which then induces the long tail latency problem. We present tt F lash as a solution to this problem. tt F lash is a “tiny-tail” flash drive (SSD) that eliminates GC-induced tail latencies by circumventing GC-blocked I/Os with four novel strategies: plane-blocking GC, rotating GC, GC-tolerant read, and GC-tolerant flush. These four strategies leverage the timely combination of modern SSD internal technologies such as powerful controllers, parity-based redundancies, and capacitor-backed RAM. Our strategies are dependent on the use of intra-plane copyback operations. Through an extensive evaluation, we show that tt F lash comes significantly close to a “no-GC” scenario. Specifically, between the 99 and 99.99th percentiles, tt F lash is only 1.0 to 2.6× slower than the no-GC case, while a base approach suffers from 5–138× GC-induced slowdowns.

Machine learning assisted OSP approach for improved QoS performance on 3D charge‐trap based SSDs

“…Some works propose to decrease the I/O performance variation and reduce the worst-case latency by garbage collection oriented optimizations. [14][15][16][17] Gugnani et al 18 propose a set of strategies for providing QoS performance guarantee with non-volatile memory express (NVMe) SSDs on virtual environment. Wu et al 2 propose a reinforcement learning-assisted I/O merging technique for improved QoS performance on SSDs.…”

Warcip

“…Garbage Collection and Write Amplification: When the number of clean blocks reduces to a certain amount, GC is triggered to generate clean pages [36]. The block chosen for GC often contains valid pages.…”

“…Because of its importance, extensive research has been reported in the literature to reduce the negative impact of GC in flash memory SSDs [14,20,34] and to cluster streaming data for data mining [2,7]. For a comprehensive review, readers are referred to [11,29,36]. In this section, we discuss prior researches that are closely related to our work.…”

Boosting performance of SSD with chip-level RAID by deferring garbage collection

“…1. Chip-level RAID within SSD provides tolerance against chip-level fault as many previous works [3,6] presented. As shown in Fig.…”

“…There exist some recent works [3,4] on optimizing the performance of user requests which are affected by GC, and they achieve good results. There is a new trend that SSDs are deployed with chip-level RAID-5 [3,5,6,7,8], to ensure fault tolerance, as MLC/TLC flash memory technology increases SSDs' capacity but sacrifices their reliability [6,9]. However, in this setting, the interference between GC and user requests becomes much more severe due to the fact that RAID-5 incurs more writes when updating parity chunks.…”