Last year at VMworld 2011 we presented one million I/O operations per second (IOPS) on a single vSphere 5 host (link). The intent was to demonstrate vSphere 5′s performance by using mutilple VMs to drive an aggregate load of one million IOPS through a single server. There has recently been some interest in driving similar I/O load through a single VM. We used a pair of Violin Memory 6616 flash memory arrays, which we connected to a two-socket HP DL380 server, for some quick experiments prior to VMworld. vSphere 5.1 was able to demonstrate high performance and I/O efficiency by exceeding one million IOPS, doing so with only a modest eight-way VM. A brief description of our configuration and results is given below.
Configuration:
Hypervisor: vSphere 5.1
Server: HP DL380 Gen8
CPU: 2 x Intel Xeon E5-2690, HyperThreading disabled
Memory: 256GB
HBAs: 5 x QLE2562
Storage: 2 x Violin Memory 6616 Flash Memory Arrays
VM: Windows Server 2008 R2, 8 vCPUs and 48GB.
Iometer Config: 4K IO size w/ 16 workers
Results:
Using the above configuration we achieved 1055896 total sustained IOPS. Check out the following short video clip from one of our latest runs.
Look out for a more thorough write-up after VMworld.

Pingback: 1 миллион IOPS с одной ВМ в vSphere 5.1 | vMind.ru
Pingback: Who needs a million IOPS for a single VM? | VCritical
Pingback: The Best Keep Getting Better – VMware vSphere 5.1 Performance | VMware vSphere Blog - VMware Blogs
Pingback: vCloud Suite – 業界で初めて Software-Defined Datacenter を実現するソリューション | Japan Cloud Infrastructure Blog - VMware Blogs
Joshua – Was there any other information added after VMWorld?
Regards
Pingback: vSphere 5.1 IOPS Performance Characterization on Flash-based Storage | VMware VROOM! Blog - VMware Blogs
Pingback: » Dr Performance: One Million IOPS Mike Laverick…
Hi Joshua,
Can you please explain how the 1 million IOPS test was done? I’ve seen the hardware specs mentioned in the blog but you guys don’t say anything about the “storage lay-out” being used during this test.
So… each of the two Violin Arrays have X amounts of LUNS carved. LUNS per array are combined together with file-system-X, and X-amount of file systems are equally divided over the (probably) four para-virtual adapters used with the test VM.
Any input will be highly appreciated!
Cheers, AJ.