Marketing fun: NetApp industry first of up to 13 million IOPS in a single rack

I’m seeing some really “out there” marketing lately, every vendor (including us) trying to find an angle that sounds exciting while not being an outright lie (most of the time).

A competitor recently claimed an industry first of up to 1.7 million (undefined type) IOPS in a single rack.

The number (which admittedly sounds solid), got me thinking. Was the “industry first” that nobody else did up to 1.7 million IOPS in a single rack?

Would that statement also be true if someone else did up to 5 million IOPS in a rack?

I think that, in the world of marketing, it would – since the faster vendor doesn’t do up to 1.7 million IOPS in a rack, they do up to 5! It’s all about standing out in some way.

Well – let’s have some fun.

I can stuff 21x EF560 systems in a single rack.

Each of those systems can do 650,000 random 4K reads at a stable 800 microseconds (since I like defining my performance stats), 600,000 random 8K reads at under 1ms, and over 300,000 random 32KB reads at under 1ms. Also each system can do 12GB/s of large block sequential reads. This is sustained I/O straight from the SSDs and not RAM cache (the I/O from cache can of course be higher but let’s not count that).

See here for the document showing some of the performance numbers.

Well – some simple math shows a standard 42U rack fully populated with EF560 will do the following:

  • 13,650,000 IOPS.
  • 252GB/s throughput.
  • Up to 548TB of usable SSD capacity using DDP protection (up to 639TB with RAID5).

Not half bad.

Doesn’t quite roll off the tongue though – industry first of up to thirteen million six hundred and fifty thousand IOPS in a single rack. 🙂

I hope rounding down to 13 million is OK with everyone.



Technorati Tags: , , ,

7 Replies to “Marketing fun: NetApp industry first of up to 13 million IOPS in a single rack”

  1. I don’t get it.. I did a quick search, and i saw IBM has a FlashSystem 840 that would fit 21 servers in a 42U with over 1 million IOPS per server.. what if I put a full rack of servers in, loaded with a software provider such as Atlantis that says they can do 1million IOPS with their software? that would be around 4 UCS chassis with 8 servers each, memory/vmware would be around 32 million IOPS..

    Maybe its just me. I don’t see as it being anything unique.

    1. I think you missed the “fun” part in the title and the entire “up to” material in the body of the text 🙂

      So we are unique in that we can do UP TO 13 million sub-millisecond 4K random IOPS in a rack – IBM (assuming they’re talking about similar IOPS characterization) would be unique in that THEY can do 21 million IOPS in a rack, etc.

      Trying to educate people how to decipher marketing material.

  2. Hey Dimitrius, hope all is well.

    That’s some fantastic numbers and kudos on publishing larger and more real-world IO sizes. I’d like to see more of that across the industry. As someone who likes diving into the details, can you share which form of RAID was used in achieving these results?

    — cheers,

    1. RIP

      I wouldn’t make that assumption. NetApp has a history of being transparent with their benchmark configuraitons in order to promote results that are achievable in the real-world. Let’s see what Dimitris shares.

      (Dimitris – sorry I got autocorrected with ‘Dimitrius’ in my first comment)


      1. Hey Vaughn,

        A few things:

        1: This was a tongue in cheek post (yet still technically accurate) to illustrate how silly that other vendor’s “industry first” marketing stuff was

        2: Rip is right, those numbers are achievable with all RAID levels and I will add are actual real-world numbers and achievable if you have beefy enough infrastructure to push the systems

        3: Notice that I did put latency numbers in the performance figures vs the block sizes. The vendor you represent talks all the time about big block performance yet doesn’t show latency for those block sizes… making the numbers pointless.



Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.