Intel has announced a boosted follow-on to its 510 SSD: the 520 Series, with more than double the IOPS performance and a top-end model with almost twice the capacity. The 2.5-inch 520 is a client or PC SSD, and is built from 25nm 2-bit MLC NAND; the 510 used a 34nm process. Like the 510, it has a 6Gb/s SATA interface, but its …
a pretty pathetic 8,000
That's still 40 times the performance of an enterprise level 15K rpm spinny thing unless you start back logging the requests to let the merge sort algorithms swap transaction rates for latency. Even so you're not likely to see one hit 500tps.
The problem with "pathetic" is most SSDs (not being sold on the China high streets with USB drives in them instead) pull well over 8K, most in the realm of 30-40k write IOPS.
Intel thinks we may as well spend those insufficient house deposits..
Can we have figures after a suitably long soak test, i.e. not fresh-from-the-box ?
Cost of SSD
If iPhones weren't soaking up most of the world's supply of chips, the price could come down.
How fast will it be after it wakes up one morning thinking its capacity is 8MB?
Won't melt in your...
server? Subliminal / sublingual message here? Trust us, won't overheat and make a mess even at speed...
(Besides, they're just working on the same theory as tools calendars and such, where each pretty-scantily-clad pretty holds an example tool in a way appealing to those having (or not having enough) tool(s). Here, they're hoping to catch you between meals, and suggesting their product will satisfy just like M&Ms - bet you can't stick in just one!)
If you are referring to those blue 'pebbles' underneath the SSD, I think they are Mentos. Only that I have never seen one with deep blue variety. Or it could be spray painted. Either M&Ms or Mentos, Intel should keep them away from kids.
Who's going to be the first (and second) to post results for SSS PTS benchmark to end all SSD benchmarks? Come on, the first will be the World Number1! Great for a press release!
Using which benchmarking tool?
Has anyone developed a piece of software for this yet?
Has anyone tried to restore an SQLServer database to an SSD array?
Testing half a dozen ATA INTEL SSDSA2M160 SSDs (latest firmware), my initial testing suggests at least double the restore time versus the same number of 10K spinning SAS disks, no matter what the RAID format being used. Some combinations/formats were 5 times worse (retesting those for verification).
It is all well and good getting great read times, but it's not a great deal of use when things go titsup on a TB or two of database that has a 6 day incremental to apply to the last full backup and I am down for a day while it gets rebuilt.
SQL Server? Your times don't sound right
@Philip Lewis: I've done a ton of testing on the OCZ Agility and IBIS drives and your numbers just don't stack up. Some figures can be found on my paper at http://www.reportingbrick.com.
I'd be interesting in you posting the hardware specifications, how you are attaching the SAS disks and how the SSD - note, the SSD's will likely be SATA connected and on say a HP controller will be limited to 1.5Gbits/second becaus of the SATA revision they use so the overall transfer speed will be significantly less that the SSD can achieve.
Back to the article, those figures are way behind what OCZ have had out for a couple of years now.
The fact that you have started quoting irrelevant hardware specs at me suggests you may not have done this sort of thing much - but I will give you the benefit of the doubt and post a single, simple simple test result.
FWIW this is all new kit, and each test is identical. Except for the trivialities of the OS (WinServer 2008R2-64) managing itself, the single variable is the data storage device.
HP G7 Blade twin x5570, 48GB RAM
P410i RAID controller (in a slot, not the built-in controller)
12 slot Storage Blade (I forget the item number)
Everything is at latest revision level.
The tests were run twice each with 6*10K 73GB SAS disks and 6*SSD devices (see previous post for item number) configured and formatted identically. Re-running of the tests produced no material differences in performance.
Test: Restore an SQLServer db with data and log files on the same array. 1 full backup save set plus 1 incremental save set.
SSD-RAID-10 = 32 minutes
SAS-RAID-10 = 7 Minutes
SSD-RAID-50 = 45 minutes
SAS-RAID-50 = 31 minutes
This is not a test to see what the fastest configuration is (I can make it go faster). I have lots of results and lots of configurations. This result is merely to demonstrate that people who think that SSDs are a silver bullet, and who busily quote electrical specifications have missed the point entirely.
FWIW, the single and low thread/queue count read speeds are at the expected levels, but recovering a multi-TB database might be some way outside our operational parameters.
I will now wander off to your link and see what you you have had to say on matters related.
Another no clue SE?
Ok, you got me - 19 years of MS SQL Server experience and 13 MVP awards aside, I've tried googling the spec of the INTEL SSD you've specified and it doesn't appear to exist even on INTEL site itself.
A quick check on the P410i specs and you'll see that SAS connected drives operate up to 6Gbits/sec (per channel or drive in english), however, SATA connected drives, which is probably what you are testing with (please show me the spec otherwise) are connected at 3Gbit/sec which is significantly short (max 264MiB/sec) of what a standard SSD is capable of - the majority are now SATA 3 (6 Gbits/sec) hence the capable 530MiBytes/sec throughput (per drive).
I hope you enjoyed my paper!
Anyway, if you would like to show us all just how almighty and experienced you are then why not blog the perfmon results and a proper spec including the correct model numbers of the drives you are using!
Slower then adata AS510S3-120GM-C
read speed 550 MB/s
write speed 510 MB/s
price 11549 yen
source: http://kakaku.com/item/K0000292178/spec/ (Japanese)
- Facebook offshores HUGE WAD OF CASH to Caymans - via Ireland
- Mexican Cobalt-60 robbers are DEAD MEN, say authorities
- Apple's spamtastic iBeacon retail alerts launch with Frisco FAIL
- Submerged Navy submarine successfully launches drone from missile tubes
- Apple sends in the bulldozers as Fruit Loop construction begins