Innovate With Global Influence


Innovate With Influence


by Steve Todd

EMC Proven

Blog powered by Typepad

Disclaimer

  • The opinions expressed here are my personal opinions. Content published here is not read or approved in advance by EMC and does not necessarily reflect the views and opinions of EMC nor does it constitute any official communication of EMC.

« John Roese CTO Roundtable at VMworld | Main | Engine Swap »

September 04, 2013

TrackBack

TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00e5500d4900883401901ef32f99970b

Listed below are links to weblogs that reference MCx: Affine Way to Core Affinity:

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

Alex

Steve, what is the underlying operating system? Linux? FreeBSD?

Steve Todd

VNX is Embedded Windows, VNXe is LINUX

INDStorage

Hi Steve,

Thanks for your very informational posts. I'm wondering if you can give some insight into a question I've pondered for some time. Obviously, the leap to Westmere chips in the VNX and now Sandy Bridge in VNX Next Gen has led to a huge increase increase in overall performance capabilities of the mid-tier array. That being said, why did it take so long to finally embrace latest generation server-class CPU's in storage array controllers? Was the performance just not needed back then (pre-flash), was it cost, or some other reason?

Pre-VNX, most CPU's that were released in storage controllers were typically not the highest speed CPU's available at the time of release, and many were CPU's designed for embedded use cases (Jasper Forest). Even though EMC has now refreshed to Sandy Bridge, many other storage arrays still haven't caught up to the Westmere chips in the Classic VNX.

Daniel Cummins

Certainly FLASH was a major factor. With previous generations the FLARE operating environment leveraged asymmetric multiprocessing which was good enough through about 4 cores. In order to leverage FLASH and to scale with the Intel roadmap we needed change our approach to multiprocessing. We could not justify this investment until the next gen VNX which leveraged much more than 4 cores.

The comments to this entry are closed.