Zbr's days.
←March→
Sun Mon Tue Wed Thu Fri Sat
            1
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31          
←2008→
Months
Jan Feb Mar
Apr May Jun
Jul Aug Sep
Oct Nov Dec

About TODO Blog RSS Old blog Projects Gallery Notes

Sun, 30 Mar 2008

To SSD or not to SSD.

Couple of days ago I talked with person, who ordered 4 high-end 128G SSD disks to create RAID for testing purposes, seek time for that devises is 0.1ms. Each one costs about $4k. His main workload is databases, i.e. random reads and writes, so we calculated that theoretically it has to be about 14 times faster than high-end scsi disks with 3.5 ms seek latency and about 100Mb/ssequential access speed in given workload for processing random data at 8-16kb chunks (usual 'page' in sql servers). Besides the fact, that putting 14 disks into mirror will be as fast as single ssd disk (theoretically), it will be 14 times more reliable and likely have smaller price, main workload is to replace RAM with SSD, not disks with SSD.

My prognosis is that SSD will be at most 2-3 times faster (if will be fater at all, since its theoretical performance advantages can be killed by FS) than SCSI disk for given workload, and as is, it is not a breakthrough technology.

If I'm wrong (it will be tested likely next week with sysbench read-write benchmark), I will buy a good bottle of whiskey for us, otherwise...

/devel/fs :: Link / Comments (5)

Simon wrote at 2008-03-31 01:15:

> since its theoretical performance advantages can be killed by FS

Oracle folks use to bypass the file systems on old fashion Unix systems, either using raw partitions or tuning the file system parameters so it doesn't get in the way.

Does the whiskey depend one whether the disks have write cache enabled or not? Enquiring mind want to know - not that I can drink whiskey these days (alas). But as someone who has benchmarked the write gains on email handling by using hardware RAID with small amounts of solid state in RAID controllers I'd be backing your colleague if these disks don't have write cache.

Anonymous wrote at 2008-03-31 02:16:

I think you might lose that bet. See http://www.nextlevelhardware.com/storage/battleship/ . If Windows can have that big of a performance improvement, given its filesystem technology, I have no doubt that Linux can.

Zbr wrote at 2008-03-31 11:31:

Well, there only single test where single ssd outperforms raptor more than 2-3 times: iometer with multiple outstanding ios, all other tests show at most 2.5 times faster results, so I should not lose :)

Simon, yes, there is possibility to bypass usual fs, but then database essentially implements own, which may or may not be tuned for ssd. As disk parameters - I do not yet know which ssd or scsi disks will be used, but pretty sure that it will be high-end vs high-end, so expect scsi has write cache and tcq.

alexr wrote at 2008-04-03 23:14:

We use SSD for some new projects. SSD can't beat SCSI disk :) in all cases where we need frequent data re-writes because of small re-write resources.

alexr wrote at 2008-04-03 23:14:

We use SSD for some new projects. SSD can't beat SCSI disk :) in all cases where we need frequent data re-writes because of small re-write resources.

Please solve this captcha to be allowed to post (need to reload in a minute): 97 + 88

Comments are closed for this story.