Review Index:
Feedback

An Inside Look at Intel's NVM Solutions Group - SSD Validation Testing

Subject: Editorial, Storage
Manufacturer: Intel

Compatibility Validation and Conclusion

Compatibility Validation:

With all of the lower level design, development, testing, and certification out of the way, the resulting SSDs need to be validated on actual hardware. An SSD may perform perfectly when connected to specific test gear running synthetic tests, but without physically plugging one into various hardware systems, you'll never know if it is fully compatible or if some corner case issue causes errors or makes the drive otherwise unreadable. Intel has expanded this operation considerably since they launched the X25-M series. Evidence of this expansion is clear, as it took nearly three months to validate the fix for the X25-M performance degradation bug discovered and reported by PC Perspective, yet subsequent fixes to issues like the rare '8MB bug' with the SSD 320 and the SMART wear level indication reporting issue with the SSD 335 took only a couple of weeks to roll out. Here's a small fraction of their fairly large validation center:

View Full Size

...and by small fraction, I mean we are talking rows of racks full of systems:

View Full Size

These systems are running every conceivable configuration. I walked past groups of servers operating every possible OS and running every piece of storage-heavy software I've ever seen or heard of. Everything from VMWare ESX to Redhat to Veritas to Windows Server / Storage Spaces, running on various hardware platforms from HP to Dell to IBM/Lenovo to Cisco.

All of the effort put in by Intel and covered over the last five pages is directed at one primary goal - making SSDs that are as reliable as possible. The proof is in the pudding as demonstrated here:

View Full Size

We knew Intel had good reliability in comparison to other manufacturers, but here are the actual figures. The return rate is higher than the actual failure rate simply because all returns are not due to actual failures. Intel tests and evaluates all returns regardless.

View Full Size

'FM3' at Intel's Folsom Campus, where the bulk of their SSD testing and validation takes place.

Conclusion:

Going into this event, I had a rough idea of what goes into the testing and evaluation of a computer component. That said, I was literally blown away by the sheer scope and scale of Intel's testing and validation efforts. While many of these efforts are geared towards their enterprise and data center parts, those gains have repeatedly trickled down into their consumer products, as evidenced by the recent SSD 730 being a consumer product based almost exactly on their DC 3500 part. This trickle-down effect is not limited to recent products either. The SSD 710 and 320 launched as an enterprise / consumer pair and were constructed identically. You can even go back to the very start, where Intel's X25-E and X25-M were an enterprise and 'mainstream' / consumer edition launched in tandem and were physically identical save the type of flash memory used in each. The big take away here is that the demanding requirements of the enterprise sector eventually make it within the grasp of the consumer, and in the case of Intel, it tends to happen almost instantaneously, and that is *very* good for the average Joe looking to buy an SSD.

March 21, 2014 | 08:16 AM - Posted by collie (not verified)

great article, super interesting read. However one this is irking me. The line is "The proof is in the EATING of the pudding" I know this is a stupid thing to bitch about but that one just gets on my tits, y'know?

March 21, 2014 | 12:17 PM - Posted by Allyn Malventano

Yes, but in the case of banana pudding, it makes you temporarily radioactive :)

March 21, 2014 | 12:03 PM - Posted by truk007

Wow! Impressive article. I love behind the scenes articles such as these. So much is involved in quality products.

March 21, 2014 | 02:12 PM - Posted by castlefox (not verified)

Allyn Malventano,

Great article, but I am wondering.... How much would that wafer be roughly worth if that stiff arm worked?

March 23, 2014 | 12:30 PM - Posted by Allyn Malventano

All depends on the yield for that particular wafer. A bad run and it wouldn't be worth much at all. Also, to know the true result of my 'stuff arm' success, check out the podcast some time :)

March 22, 2014 | 05:24 AM - Posted by Anonymous (not verified)

Hi

The "X25-M performance degradation bug discovered and reported by PC Perspective" was not discovered by PC Perspective ;-)

February 13, 2009

http://www.pcworld.fr/stockage/tests,ssd-intel-x25-m-80-go-une-bombe-pro...

<= September 8, 2008

http://www.hardware.fr/articles/731-6/supertalent-intel-performances-var...

<= September 8, 2009

In fact it doesn't take 3 month to fix this problem but ... 8 month ! ;-)

March 23, 2014 | 12:28 PM - Posted by Allyn Malventano

Interesting. We hadn't seen that piece, and neither had Intel apparently, as in the communications about their first firmware update, they credited us with its discovery. 

April 5, 2014 | 07:03 PM - Posted by Jimmy Jim (not verified)

Excellent article Allyn!

April 5, 2014 | 07:03 PM - Posted by Jimmy Jim (not verified)

Excellent article Allyn!

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

By submitting this form, you accept the Mollom privacy policy.