How to store a petabyte?

Ryan Byrd ryanbyrd at gmail.com
Tue Feb 20 16:47:08 MST 2007


Let's say I need to store a petabyte of data. I need fast access (tape/DVDs
aren't fast enough) and redundancy like RAID.

How should I do it? Initially the data will be in large (several MB) binary
objects and could be stored as files, but eventually, it will need to be
placed into a relational database like Oracle.

Let's say I have access to racks that are 44U tall. I've listed several very
different scenarios. Which is best? Is there a better way? How would *you*
store 1PB? What do you/your company currently use to store your large
datasets? What kind of drives should I use? Do SCSI drives last longer than
SATA? What about IDE or SAS? Do higher RPM drives have a shorter mean time
to failure?

Scenario 1:
----------------
EMC CLARiiON CX300 PSI w/ Fiber channel
14x300GB ultrawide SCSI
$40K
3.9TB
5U
8 per rack
31.2TB per rack
$320K per rack
33 needed for 1PB
$10.6 million

Scenario 2:
----------------
Dell PowerVault MD3000
14x146GB SAS 10K drives, +1 hot spare
1.7TB usable w/ RAID 5
3U
$16K each
14 per rack
$224K per rack
23.8TB per rack
43 racks needed for 1PB
$9.6 million

Scenario 3:
----------------
HP ProLiant DL320s Server
12 250GB SATA 7.2k drives
2.7GB with RAID 5
3U
$5.8K
14 per rack
$81.2K per rack
37.8TB per rack
27 racks needed for 1PB
$2.2 million

Scenario 4:
----------------
HP ProLiant DL320 G4
2 SATA 500GB 7.2k
490GB usable with software RAID 1
1U
$2.8K each
44 per rack
21.56TB per rack
$123.2K per rack
47.5 racks needed for 1PB
$5.9 million



More information about the PLUG mailing list