A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2016; you can also visit the original URL.
The file type is application/pdf
.
Filters
Preemptible I/O Scheduling of Garbage Collection for Solid State Drives
2013
IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
In this paper, we examine the GC process and propose a semi-preemptible GC scheme that allows GC processing to be preempted while pending I/O requests in the queue are serviced. ...
Write-dominant workloads show up to a 66.56% improvement in average response time with a 83.30% reduced variance in response time compared to the non-preemptible GC scheme. ...
Also this work was also partially sponsored through Korea Ministry of Knowledge Economy grant (No. 10037244). ...
doi:10.1109/tcad.2012.2227479
fatcat:iyvjmjef7fd7lg43ua2y4nkh24
Tiny-Tail Flash: Near-Perfect Elimination of Garbage Collection Tail Latencies in NAND SSDs
2017
USENIX Conference on File and Storage Technologies
TTFLASH is a "tiny-tail" flash drive (SSD) that eliminates GC-induced tail latencies by circumventing GCblocked I/Os with four novel strategies: plane-blocking GC, rotating GC, GC-tolerant read, and GC-tolerant ...
We show that TTFLASH comes significantly close to a "no-GC" scenario. ...
O schedulers [49] , and disk/SSD hardware-level defects [26, 27, 30 ]. ...
dblp:conf/fast/YanLHTSCG17
fatcat:kw2kwv5pczf6vn36iqofp2khvq
SplitZNS: Towards an Efficient LSM-tree on Zoned Namespace SSDs
2023
ACM Transactions on Architecture and Code Optimization (TACO)
ZNS exposes erase blocks in SSD as append-only zones, enabling the LSM-tree to gain awareness of the physical layout of data. ...
In this paper, we present SplitZNS, which introduces small zones by tweaking the zone-to-chip mapping to maximize GC efficiency for LSM-tree on ZNS SSDs. ...
This is because a smaller value size incurs more CPU overhead, making the LSM-tree less I/O intensive. ...
doi:10.1145/3608476
fatcat:akqxqnvkrjewhjukp3lf4dwdxm
LinnOS: Predictability on Unpredictable Flash Storage with a Light Neural Network
2020
USENIX Symposium on Operating Systems Design and Implementation
Our evaluation shows that, compared to hedging and heuristicbased methods, LinnOS improves the average I/O latencies by 9.6-79.6% with 87-97% inference accuracy and 4-6µs inference overhead for each I/ ...
O, demonstrating that it is possible to incorporate machine learning inside operating systems for real-time decision-making. ...
To the best of our knowledge, there is no existing learning approach for I/O scheduling that supports such fine-grained learning due to the challenges of achieving per-I/O accuracy and fast online inference ...
dblp:conf/osdi/HaoTLHHG20
fatcat:3dzlbnmdcfhyfjhlhelz77guq4
Caladan: Mitigating Interference at Microsecond Timescales
2020
USENIX Symposium on Operating Systems Design and Implementation
When colocating memcached with a best-effort, garbage-collected workload, Caladan outperforms Parties, a state-of-the-art resource partitioning system, by 11,000×, reducing tail latency from 580 ms to ...
Unfortunately, partitioning-based systems fail to react quickly enough to keep up with these changes, resulting in extreme spikes in latency and lost opportunities to increase CPU utilization. ...
(a) THRESH_QD allows an operator to achieve better tail latencies at the expense of BE throughput. (b) THRESH_HT reins in the latency of long requests, but setting it too low reduces BE throughput. ...
dblp:conf/osdi/FriedROB20
fatcat:ks6wxbodlbhrdmnnxxxcdcd7f4
A Survey on Tiering and Caching in High-Performance Storage Systems
[article]
2019
arXiv
pre-print
In software, caching and tiering are long-established concepts for handling file operations and moving data automatically within such a storage network and manage data backup in low-cost media. ...
In this survey, we discuss some recent pieces of research that have been done to improve high-performance storage systems with caching and tiering techniques. ...
A study [29] on using NVM as an I/O cache for SSD or HDD reveals that the current I/O caching solution cannot fully benefit from the low-latency and high-throughput of NVM. ...
arXiv:1904.11560v1
fatcat:e752fsvuzbcxtmqjxg4ezlptku
HybridStore: A Cost-Efficient, High-Performance Storage System Combining SSDs and HDDs
2011
2011 IEEE 19th Annual International Symposium on Modelling, Analysis, and Simulation of Computer and Telecommunication Systems
collection (GC) induced by high intensity of random writes). ...
Given these trade-offs between HDDs and SSDs in terms of cost, performance, and lifetime, the current consensus among several storage experts is to view SSDs not as a replacement for HDD but rather as ...
As we increase the I/O intensity, we observe the need for MLC SSDs to satisfy the bandwidth requirements with increased I/O intensity. ...
doi:10.1109/mascots.2011.64
dblp:conf/mascots/KimGUBS11
fatcat:sb4z7soyybal7kwox6b4j6npju
Improving flash write performance by using update frequency
2013
Proceedings of the VLDB Endowment
In this paper, we show how to design FTLs that are more efficient by using the I/O write skew to guide data placement on flash memory. ...
a given I/O workload. ...
Latency-wise, garbage collection scheduling has a significant impact on the I/O latency, however, it is orthogonal to the problem of data placement. ...
doi:10.14778/2536360.2536372
fatcat:dhzuhr2jmfam7fpzoeljf5n27u
A Temporal Locality-Aware Page-Mapped Flash Translation Layer
2013
Journal of Computer Science and Technology
For example, a predominantly random-write dominant I/O trace from an OLTP application running at a large financial institution shows a 78% improvement in average response time (due to a 3-fold reduction ...
The poor performance of random writes has been a cause of major concern which needs to be addressed to better utilize the potential of flash in enterprise-scale environments. ...
However, for TPC-H, it exhibits a long tail primarily because of the expensive full merges and the consequent high latencies seen by requests in the I/O driver queue. ...
doi:10.1007/s11390-013-1395-4
fatcat:izcp7dstireqbcngjt4uwjq6em
Architectural Techniques for Improving NAND Flash Memory Reliability
[article]
2018
arXiv
pre-print
Raw bit errors are common in NAND flash memory and will increase in the future. These errors reduce flash reliability and limit the lifetime of a flash memory device. ...
effect to mitigate retention errors in 3D NAND. ...
Scheduling Requests The controller receives I/O requests over a host controller interface (shown as Host Interface in Figure 2 .1b), which consists of a system I/O bus and the protocol used to communicate ...
arXiv:1808.04016v1
fatcat:fotned4yajc2xmaoezwjdrgypu
ICE: Managing cold state for big data applications
2016
2016 IEEE 32nd International Conference on Data Engineering (ICDE)
However, these M3 applications require the SPE to maintain massive amounts of state in memory, leading to resource usage skew: memory is scarce and over-utilized, whereas CPU and I/O are under-utilized ...
The use of big data in a business revolves around a monitor-mine-manage (M3) loop: data is monitored in real-time, while mined insights are used to manage the business and derive value. ...
This optimization is able to reduce disk write I/O by 64%. ...
doi:10.1109/icde.2016.7498262
dblp:conf/icde/ChandramouliLC16
fatcat:mzm3bppvbjgbjmyromqoklcnpa
Semeru: A Memory-Disaggregated Managed Runtime
2020
USENIX Symposium on Operating Systems Design and Implementation
In contrast, programs written in managed languages are subject to periodic garbage collection (GC), which is a typical graph workload with poor locality. ...
modifications; (2) a distributed GC, which offloads object tracing to memory servers so that tracing is performed closer to data; and (3) a swap system in the OS kernel that works with the runtime to ...
We are grateful to our shepherd Yiying Zhang for her feedback, helping us improve the paper substantially. This work is supported by NSF grants CCF- ...
dblp:conf/osdi/WangMLLRNBNKX20
fatcat:mleexavtujcwjjovvu2ffesdtu
The RAMCloud Storage System
2015
ACM Transactions on Computer Systems
RAMCloud is a storage system that provides low-latency access to large-scale datasets. To achieve low latency, RAMCloud stores all data in DRAM at all times. ...
RAMCloud uses a unique two-level approach to log cleaning, which maximizes DRAM space utilization while minimizing I/O bandwidth requirements for secondary storage. Latency. ...
without waiting for the I/O to complete. ...
doi:10.1145/2806887
fatcat:fg3r5yahbjhxhcor6m2w2q6bxy
DAG to reduce the E2E latency and cost. ...
invocations of a function in one VM to improve resource sharing among the parallel workers to reduce skew. (3) Resource Allocation assigns the right VM size to each function or function bundle in the ...
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the sponsors. ...
doi:10.1145/3530892
fatcat:an4t5ybn6rbbdh34sxv7gpzyi4
LLAMA
2013
Proceedings of the VLDB Endowment
To demonstrate LLAMA's suitability, we tailored our latch-free Bw-tree implementation to use LLAMA. The Bw-tree is a B-tree style index. ...
SL uses the same mapping table to cope with page location changes produced by log structuring on every page flush. ...
LLAMA is unique in a number of ways. ...
doi:10.14778/2536206.2536215
fatcat:4hqukk6qczer3htcynceakurpi
« Previous
Showing results 1 — 15 out of 30 results