Review Index:
Feedback

Intel SSD DC P3520 2TB Review - 3D NAND = $0.50/GB!

Subject: Storage
Manufacturer: Intel

Latency Percentile, IO Percentile, and Saturated (High-Res) QoS

Required reading (some additional context for those unfamiliar with our Percentile testing):

Intro to PACED workloads - 'It's not how fast you go, it's how well you go fast!'

Since this is a solo review of the P3520, with no direct comparisons to other products, I've shelved our PACED Quality of Service (QoS) testing in favor of directly evaluating the QoS figures stated in Intel's product specification. I will also be introducing a novel way of presenting the QoS "9's" data which is derived from IO percentile, but before we get into that, I figured this review would be a good opportunity to take a given SSD and lay out the three presentation methods we now have at our disposal. I'll start with Latency (total time) Percentile, shift to IO Percentile, and finally we will rework that data into our new High Resolution QoS plot, which will allow easy comparison to the specification (marked with X's as we did earlier in this review).

Latency Percentile

View Full Size

View Full Size

Latency Percentile is a translation of the latency distribution that takes into account the time spent servicing the IOs, meaning the above plots are showing the percentage of the total run time spent on those IOs. The results are effectively weighed by latency, where longer IOs have a larger impact on the percentile. These results are *not* used for QoS calculations, since QoS assumes the IOs are all independent, meaning if one IO stalls, the rest will just keep on going without waiting for it.

IO Percentile

View Full Size

View Full Size

Looking at the pure IO Percentile spread, we note some OS RAM cache hits accounting for a small percentage of reads. Those were minimized in Latency Percentile because those IOs were so much faster than the bulk and took a very small fraction of the total time to complete.

Quality of Service (QoS)

QoS is specified in percentages (99.9%, 99.99%, 99.999%), and uniquely spoken (‘three nines’, ‘four nines’, ‘five nines’). It corresponds to the latency seen at the top 99.x% of all recorded IOs in a run. Enterprise IT managers and system builders care about varying levels of 9's because those long latencies lead to potential timeouts for time-sensitive operations, and increasing the 9's is how they quantify more stringent QoS requirements. Note that these comparative results are derived from IO Percentile data and *not* from Latency Percentile data.

If you have a hard time wrapping your head around the 9's thing, It may be easier to flip things around and think about it from the standpoint of the remaining longest-latency IO's that haven't been accounted for as the plot progresses. As an example, the 99.9% line near the center of the vertical axis represents the top 10% of the top 1% (0.1%) of all recorded IOs, where 'top' means those IOs of the longest latency.

These plots are tricky to make, as they are effectively an inverse log scale. Each major increment up from the zero axis corresponds to the top 90%, and the next increment after that shows the top 90% *of that previous value*, meaning it's an asymptotic scale which will never reach 100%. The plots below essentially take the top portion of the IO Percentile results and spread them out, exponentially zooming in on the results as they approach 100%.

First the specs:

View Full Size

...and now the results (X'd as appropriate):

View Full Size

View Full Size

QoS was within spec in nearly all cases. There were some slight oddities noted with reads, but still well within a small margin of error considering my testing is done under Windows while Intel's rating is derived using Linux (and with a completely different test suite).

The high-resolution latency distribution binning employed back when I started this whole percentile thing was actually done as a means to get more accurate QoS figures, and with High-Resolution QoS now a fruit of that labor, the circle is now complete. In a further twist, these very results were fed back to Intel and played a part in tweaking their QoS specification for the P3520! High Resolution QoS for the win!


August 25, 2016 | 02:06 PM - Posted by Anonymous (not verified)

$0.50/GB is considered good? Was this article written in 2005?

August 25, 2016 | 02:26 PM - Posted by Vitagra

For pci-e ssds, that is considered good.

August 25, 2016 | 03:46 PM - Posted by Nimrod (not verified)

Yeah, for SATA SSDs anything <0.25/GB is pretty good, this is about twice that but you're also getting around twice the speeds.
Too expensive for me personally, but not unreasonable IMO.

August 25, 2016 | 04:32 PM - Posted by Allyn Malventano

Intel enterprise SSDs didn't launch until 2008, and did so at >$10/GB (>20x the cost).

August 25, 2016 | 08:00 PM - Posted by Anonymous (not verified)

That's good progress, so they should begin to be viable around 2024

August 25, 2016 | 09:31 PM - Posted by Allyn Malventano

SSD market share has doubled for the past two years. It's expected to surpass HDD a lot sooner than 2024.

January 31, 2017 | 10:51 AM - Posted by John H (not verified)

in 2005 SSDs would be more like $50/GB :)

August 26, 2016 | 01:46 AM - Posted by Anonymous (not verified)

For that terrible 0.7 DWPD/5 years, I would take 750 over this thing any day, performance wise it's not even close to P3700/750.

August 26, 2016 | 04:53 PM - Posted by Allyn Malventano

Performance is no comparison, obviously. The point of this drive is cost, which is a fraction of all parts you mentioned.

August 26, 2016 | 05:53 AM - Posted by Jann5s

Allyn, thank you, I really like the depth of your reviews, I'm actually learning stuff!

August 26, 2016 | 09:44 AM - Posted by Anonymous (not verified)

I do not find any mention of capacitor for power loss writes. It's a feature on which I place great importance.

August 26, 2016 | 04:52 PM - Posted by Allyn Malventano

Intel has among the highest, if not *the* highest power loss testing / qualification / reliability in the industry. It wasn't mentioned specifically because at this point it's just a given for their products. Here's a blurb from one of their product briefings:

Intel. Intel® Datacenter Drives provide robust Power Loss Imminent (PLI) circuitry that helps to protect inflight data in the event of power loss. Intel drives monitor the health of the PLI circuitry via a Self Cap Test using SMART attributes.  Samsung PM853T and SM843T drives were checked for capabilities and flags.  No PLI monitoring capabilities (e.g. SMART Attributes) were listed in the Samsung drive specification sheet.  Additionally, the drives were tested by powering off a drive and removing one electrolytic (or any other type) capacitor. The drives were then powered up to recollect SMART attribute data to determine is the cap test detected the removal of the capacitor.  The Samsung drives did not detect capacitor removal. 

They also bombard their drives with radiation (from an accelerator) until they hang, restart them, and ensure no data was corrupted. Their testing is pretty crazy, and that's why their products typically run higher in cost compared to others, but you get what you pay for.

 

August 27, 2016 | 03:45 AM - Posted by Anonymous (not verified)

Many think inflight data protection only as a safety issue, but it is also a significant performance issue. Without inflight data protection, use of inflight data must be turned off in the OS (it may be called something like write cache) to avoid data corruption in case of power failure, which in turn significantly lowers write speed.

So the point of inflight data protection or the lack of it should be hammered home in every review until it gets the warranted attention.

August 28, 2016 | 02:07 AM - Posted by Allyn Malventano

There are lots of layers of what would/could be considered 'in-flight'. Even with all caching disabled, the mere fact that writes are queued could be considered so, as they are technically buffered by the kernel. To strip all the way down to zero buffering would reduce the performance of *most* SSDs to painful levels, as you'd have to limit to QD=1 and disable all OS buffers.

This protection, as defined by SSD makers, is a guarantee that the data that has been received by the controller at the point of power loss will be retained and available at next power up. Host / OS-side buffers will naturally not be included here.

August 26, 2016 | 11:15 AM - Posted by dstanding (not verified)

Very excited about P3520 especially in U.2 2.5" format. This kind of pricing should really increase the viability (economically speaking) of big top-of-rack all flash arrays.

Not sure if you mentioned in the review but has Intel made any mention of dual-port U.2 version?

August 26, 2016 | 04:53 PM - Posted by Allyn Malventano

No mention of dual port for this one, but I'd guess once 3D rolls out to other models in their lineup, it will include dual port.

September 1, 2016 | 10:03 PM - Posted by So.... (not verified)

So, let me make sure I understand. This SSD is not tested against any other product, yet receives an editors choice. I smell something.

September 10, 2016 | 03:55 PM - Posted by Allyn Malventano

What you smell is no other products competing at this low of a cost/GB. Other companies are welcome to sample us their competing products (we ask them often).

September 18, 2016 | 04:36 PM - Posted by Anonymous (not verified)

It was pretty well-explained why...

September 6, 2016 | 08:40 AM - Posted by "; echo killer (not verified)

what about raid 0 on 4 of these

September 12, 2016 | 09:47 AM - Posted by Guido (not verified)

We are thinking of using the P3520 or P3500 in Supermicro 48 bay nvme server. P3500 might be quicker but probably these will already move the bottleneck to the interface... Will have a look if you benchmarked the p3500 before...

November 1, 2016 | 05:35 PM - Posted by Stan (not verified)

Going to try out three of the 1.2TB P3520's for the hot tier in a three node hyperconverged environment. It'd be interesting to know what sort of benchmark would be relevant for comparison purposes on that kind of platform, since the workload mix could look like practically anything.

November 1, 2016 | 06:02 PM - Posted by Jeremy Hellstrom

Yes it would, trying to set up benchmarks simulating that kind of environment is not simple.  Let us know how it goes as it could be very interesting.

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.