Characterizing and Modeling Reliability of Declustered RAID for HPC Storage Systems

Qiao, Zhi; Liang, Shuwen; Fu, Song; Chen, Hsing-Bung; Settlemyer, Bradley W.

doi:10.1109/dsn-industry.2019.00011

Cited by 7 publications

(5 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…At inter-node level, stacking network switch allowing redundant network links using Link Aggregation Control Protocol (LACP) is recommended, while at storage level, Redundant Arrays of Inexpensive Disks (RAID, see, e.g. Patterson and Hennessy, 2016) and evolutions such as Declustered RAID, see Qiao et al (2019), should feature.…”

Section: Resilience Methodologiesmentioning

confidence: 99%

Resilience and fault tolerance in high-performance computing for numerical weather and climate prediction

Benacchio

Bonaventura

Altenbernd

et al. 2021

The International Journal of High Performance Computing Applica

View full text Add to dashboard Cite

Progress in numerical weather and climate prediction accuracy greatly depends on the growth of the available computing power. As the number of cores in top computing facilities pushes into the millions, increased average frequency of hardware and software failures forces users to review their algorithms and systems in order to protect simulations from breakdown. This report surveys hardware, application-level and algorithm-level resilience approaches of particular relevance to time-critical numerical weather and climate prediction systems. A selection of applicable existing strategies is analysed, featuring interpolation-restart and compressed checkpointing for the numerical schemes, in-memory checkpointing, user-level failure mitigation and backup-based methods for the systems. Numerical examples showcase the performance of the techniques in addressing faults, with particular emphasis on iterative solvers for linear systems, a staple of atmospheric fluid flow solvers. The potential impact of these strategies is discussed in relation to current development of numerical weather prediction algorithms and systems towards the exascale. Trade-offs between performance, efficiency and effectiveness of resiliency strategies are analysed and some recommendations outlined for future developments.

show abstract

Section: Resilience Methodologiesmentioning

confidence: 99%

Resilience and fault tolerance in high-performance computing for numerical weather and climate prediction

Benacchio

Bonaventura

Altenbernd

et al. 2021

The International Journal of High Performance Computing Applica

View full text Add to dashboard Cite

show abstract

“…( 12). The primitive polynomial of GF( 2) is obtained using f(x); then, GF (8) in Eq. ( 9) is used to obtain p(x).…”

Section: Decodingmentioning

confidence: 99%

“…Use f(x) in Eq. ( 13) to replace all the parameter calculations in GF (8). The RS-RAID 6 calculation will be simplified to c 1 and c 2 multiplication.…”

Section: Decodingmentioning

confidence: 99%

“…Then, c 2 can be transformed to α 25 . By the same derivation, we can calculate each dn in GF(2) for GF (8).…”

Section: Decodingmentioning

confidence: 99%

“…For example, EVENODD and RDP both have to update three (on average) parity blocks whenever a data block is changed; this number is 1.5 times the theoretical lower bound. (7)(8)(9)(10) RAID systems have received increasing attention with the ever increasing probability of multiple disk failures. (11,12) RAID-based architectures are also used in clusters and large-scale storage systems.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Efficient Reed–Solomon Redundant Arrays of Independent Disks 6 Exclusive OR Accelerator with Scaling Scheme

Yue,

Wul,

Hou

et al. 2023

Sensors and Materials

View full text Add to dashboard Cite

An efficient Reed-Solomon redundant arrays of independent disks 6 (RS-RAID 6) exclusive OR (XOR) accelerator with a scaling scheme is proposed in this paper. The accelerator first considers the hardware working characteristic to reduce the complexity of codec computation. To increase efficiency, the parameter calculation of the 8-bit Galois field [GF(8)] is converted to a GF(2) calculation, where two additions, one multiplication, and one division are optimized in every 8 bits of calculation. A new scaling scheme is also proposed to guarantee the input/output (I/O) balance. The proposed Reed-Solomon (RS) codec hardware architecture is fully verified at the Real Time Logistics (RTL) level and synthesized with a GF 12 nm technology library to compare the gate cost and input/output operations per second (IOPS) performance with those of other classic solutions. Separate simulations are performed for different codec algorithms and scaling schemes. The IOPS performance of the proposed RS-RAID 6 XOR accelerator is improved by a factor of 2-3 at the cost of 47% more gates.

show abstract

Applying SDN based data network on HPC Big Data Computing – Design, Implementation, and Evaluation

Chen

Qiao

2019

2019 IEEE International Conference on Big Data (Big Data)

View full text Add to dashboard Cite

Characterizing and Modeling Reliability of Declustered RAID for HPC Storage Systems

Cited by 7 publications

References 5 publications

Resilience and fault tolerance in high-performance computing for numerical weather and climate prediction

Resilience and fault tolerance in high-performance computing for numerical weather and climate prediction

Efficient Reed–Solomon Redundant Arrays of Independent Disks 6 Exclusive OR Accelerator with Scaling Scheme

Applying SDN based data network on HPC Big Data Computing – Design, Implementation, and Evaluation

Contact Info

Product

Resources

About