3
Performance of Direct Attached Disk Subsystems Roger D. Chamberlain*†, Berkley Shands† * Exegy Inc., St. Louis, MO Dept. of Computer Science and Engineering, Washington University in St. Louis

Performance of Direct Attached Disk Subsystems

  • Upload
    artie

  • View
    28

  • Download
    0

Embed Size (px)

DESCRIPTION

Performance of Direct Attached Disk Subsystems. Roger D. Chamberlain*†, Berkley Shands† * Exegy Inc., St. Louis, MO † Dept. of Computer Science and Engineering, Washington University in St. Louis. Experimental Systems. Platform 1: single 3U enclosure 16 drives @ 500 GB 2 controllers - PowerPoint PPT Presentation

Citation preview

Page 1: Performance of Direct Attached Disk Subsystems

Performance of Direct Attached Disk Subsystems

Roger D. Chamberlain*†, Berkley Shands†* Exegy Inc., St. Louis, MO

†Dept. of Computer Science and Engineering, Washington University in St. Louis

Page 2: Performance of Direct Attached Disk Subsystems

Experimental Systems

diskco ntro ller

U niwide enclo sure

S A S switch

S A S switch

16 disks4

chan.

4chan.

16 disks

X to re enclo sure

X to re enclo sure

pro cs

FP G A

P C Ie

P C I-X

Platform 1:• single 3U enclosure• 16 drives @ 500 GB• 2 controllers• 8 TB of storage

Platform 2:• three 3U enclosures• 32 drives• 1 controller• 2 SAS expanders• 13 TB of storage

Page 3: Performance of Direct Attached Disk Subsystems

Empirical Performance

Read throughput as a functionof minimum file size• xfs and ext3 file systems• platform 2• 8 logical drives, 4 disks each

Write throughput as a functionof position on the disk• xfs file system• platform 1• 4 logical drives, 4 disks each

0

100

200

300

400

500

600

700

800

900

1000

0 32 KB 64 KB 128 KB 256 KB 512 KB 1 MB 2 MB 4 MB

Minimum file size

Rea

d t

hro

ug

hp

ut

(MB

/s)

xfs ext3

0

200

400

600

800

1000

1200

0 0.2 0.4 0.6 0.8 1

Position on disk

Wri

te r

ate

(M

B/s

)