Proceedings of the 11th ACM Symposium on Cloud Computing 2020
DOI: 10.1145/3419111.3421290
|View full text |Cite
|
Sign up to set email alerts
|

High availability in cheap distributed key value storage

Abstract: Memory-based storage currently offers the highest-performance distributed storage, keeping the primary copy of all data in DRAM. Recent advances in non-volatile main memory (NVMM) technologies promise latency similar to DRAM at reduced cost and energy, but will make providing high availability more challenging. Previous approaches to failure recovery involve maintaining multiple identical replicas or relying on fast offline restoration of data from backup replicas stored on SSD. Unfortunately, NVMM's combinati… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
23
0

Year Published

2022
2022
2023
2023

Publication Types

Select...
4
1

Relationship

0
5

Authors

Journals

citations
Cited by 5 publications
(23 citation statements)
references
References 17 publications
0
23
0
Order By: Relevance
“…However, on the multicore processor architecture, each logical core can run a host thread, and each host thread can carry a user-mode thread pool. The lack of a unified provisioning and management mechanism between multiple usermode thread pools inevitably restricts the efficiency and performance improvement of user-mode thread pools on multicore processor architecture [14,15,16,17] .…”
Section: Related Workmentioning
confidence: 99%
“…However, on the multicore processor architecture, each logical core can run a host thread, and each host thread can carry a user-mode thread pool. The lack of a unified provisioning and management mechanism between multiple usermode thread pools inevitably restricts the efficiency and performance improvement of user-mode thread pools on multicore processor architecture [14,15,16,17] .…”
Section: Related Workmentioning
confidence: 99%
“…It can have independent register contexts and execution stacks so that a single kernel thread can process multiple tasks at the same time without the need for switching between user mode and kernel mode to reduce system loss. Compared to the traditional thread running model, the coroutines are more controllable, lightweight, fast, and efficient with more obvious advantages [5,6,7] .…”
Section: Introductionmentioning
confidence: 99%
“…Raumer et al [127] offloaded the cryptographic function of a virtual private network gateway into commodity NICs, for increased performance. MICA [128] encodes database keys into the source port field of User Datagram Protocol (UDP) packets' header, which are classified by Intel NICs [103] in hardware to provide high performance inmemory key-value storage. Similarly, SwitchKV [129] offloads key-value stores into OpenFlow switches, which typically provide greater classification capabilities and more rule capacity than NICs.…”
Section: Hardware Offloadingmentioning
confidence: 99%
“…Figure 7.11 shows the latencies obtained from these experiments. Four frame sizes (64,128,256, and 1500 bytes) were used to inject 100,000 frames per frame size. The latency per frame was measured and plotted as boxplots.…”
Section: Performance Of Internal Stagesmentioning
confidence: 99%
See 1 more Smart Citation