TY - GEN
T1 - Enabling cost-effective flash based caching with an array of commodity SSDs
AU - Oh, Yongseok
AU - Lee, Eunjae
AU - Hyun, Choulseung
AU - Choi, Jongmoo
AU - Lee, Donghee
AU - Noh, Sam H.
N1 - Publisher Copyright:
© 2015 ACM.
PY - 2015/11/24
Y1 - 2015/11/24
N2 - SSD based cache solutions are being widely utilized to im- prove performance in network storage systems. With a goal of providing a cost-effective, high performing SSD cache so- lution, we propose a new caching solution called SRC (SSD RAID as a Cache) for an array of commodity SSDs. In designing SRC, we borrow both the well-known RAID tech- nique and the log-structured approach and adopt them into the cache layer. In so doing, we explore a wide variety of de- sign choices such as flush issue frequency, write units, form- ing stripes without parity, and garbage collection through copying rather than destaging that become possible as we make use of RAID and a log-structured approach at the cache level. Using an implementation in Linux under the Device Mapper framework, we quantitatively present and analyze results of the design space options that we consid- ered in our design. Our experiments using realistic work- load traces show that SRC performs at least 2 times better in terms of throughput than existing open source solutions. We also consider cost-effectiveness of SRC with a variety of SSD products. In particular, we compare SRC config- ured with MLC and TLC SATA SSDs and a single high-end NVMe SSD. We find that SRC configured as RAID-5 with low-cost MLC and TLC SATA SSDs generally outperforms that configured with a single high-end SSD in terms of both performance and lifetime per dollars spent.
AB - SSD based cache solutions are being widely utilized to im- prove performance in network storage systems. With a goal of providing a cost-effective, high performing SSD cache so- lution, we propose a new caching solution called SRC (SSD RAID as a Cache) for an array of commodity SSDs. In designing SRC, we borrow both the well-known RAID tech- nique and the log-structured approach and adopt them into the cache layer. In so doing, we explore a wide variety of de- sign choices such as flush issue frequency, write units, form- ing stripes without parity, and garbage collection through copying rather than destaging that become possible as we make use of RAID and a log-structured approach at the cache level. Using an implementation in Linux under the Device Mapper framework, we quantitatively present and analyze results of the design space options that we consid- ered in our design. Our experiments using realistic work- load traces show that SRC performs at least 2 times better in terms of throughput than existing open source solutions. We also consider cost-effectiveness of SRC with a variety of SSD products. In particular, we compare SRC config- ured with MLC and TLC SATA SSDs and a single high-end NVMe SSD. We find that SRC configured as RAID-5 with low-cost MLC and TLC SATA SSDs generally outperforms that configured with a single high-end SSD in terms of both performance and lifetime per dollars spent.
KW - Cost-effective storage
KW - Flash based caching
KW - Log-structured caching layout
KW - SSD array
UR - http://www.scopus.com/inward/record.url?scp=84967144628&partnerID=8YFLogxK
U2 - 10.1145/2814576.2814814
DO - 10.1145/2814576.2814814
M3 - Conference contribution
AN - SCOPUS:84967144628
T3 - Middleware 2015 - Proceedings of the 16th Annual Middleware Conference
SP - 63
EP - 74
BT - Middleware 2015 - Proceedings of the 16th Annual Middleware Conference
PB - Association for Computing Machinery, Inc
T2 - 16th International Middleware Conference, Middleware 2015
Y2 - 7 December 2015 through 11 December 2015
ER -