Virtualization to the Rescue

Dave Graham

Subscribe to Dave Graham: eMailAlertsEmail Alerts
Get Dave Graham via: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn

A color-enhanced version of an aerial photogra...
Image via Wikipedia

It’s a rather dark and snowy Tuesday in Hopkinton today but, despite the craptastic weather, there’s some good news floating about.  EMC’s Cloud Infrastructure Group has released Atmos version 1.3 which is a significant overhaul of the Atmos code previously released.  So, what’s new, what’s cool, and why even bother?

Out with the Old, in with the New!

Several changes are occuring within the Atmos construct: hardware and software.  I’ll start with the hardware as this is the easiest to digest.

Hardware changes

The biggest change on the Atmos hardware side is the shift away from Intel Harpertown processors (5400 series) and their egregiously hot FB-DIMMs to the new Intel Nehalem (5500 series) processors.  This shift gains better IPC, a more optimized compiling platform, as well as laying the groundwork for even further code optimizations as we integrate further into Intel’s future architecture designs.

Another change is the addition of 2Tb “green” harddrives.  These drives spin at 5400rpm and consume less power while offering roughly double the formatted capacity of the 1Tb 7200rpm solution available currently.   In a rack instantiation, this effectively doubles the given capacity of each frame (i.e. our WS-360 moves from 360Tb raw to 720Tb raw) and again sets the stage for future integration and design work.

Finally, each of the Atmos nodes themselves receive some updated hardware configurations courtesy of the updated server architecture (more GigE ports per node, etc.)

Software changes

Seeing as how Atmos is software, any changes to the stack are pretty important.  There’s a of significant update here that are worth mentioning:  GeoProtect.  So what exactly is GeoProtect?


What is GeoProtect?

GeoProtect is a Reed-Solomon derivative in which data is divided into multiple sections and distributed, by policy, throughout an Atmos Resource Management Group (RMG).  This can be distributed based on locality (e.g. on premise) or geographically across RMGs (e.g. Boston, London, and Dubai).  The GeoProtect policy can be scaled to distribute more segments (e.g. greater protection with more consumed space) or fewer segments (e.g. protection with less consumed space than 1:1 object replicas).  All in all, a good move for customers looking for more flexibility for their data protection.

So, where do we go from here?

As you may or may not know, I work within the Atmos product group as a Systems Engineer with a specific focus on virtualization and Atmos.  Contrary to popular belief, this isn’t about shoe-horning Atmos into VMware (or ANY hypervisor, for that matter) but more about looking at how virtualization complements enterprise strategies for both private AND hybrid clouds (special shout to @randybias on the “hybrid” moniker ;) ).  That being said, there’s little that can’t be done when you couple Atmos’ policy and engines to standard grade storage solutions.  I’ll leave that alone for the time being.

While I don’t have quite the amount of clout needed to declare 2010 as the year of the cloud, I’m going to go ahead and say that 2010 is the year of EMC’s Cloud.  We’ve got a tremendous road ahead and I’m excited about the journey to get to cloud maturity!

Comments and questions are welcome!

EDIT: press release is here



Reblog this post [with Zemanta]

Read the original blog entry...

More Stories By Dave Graham

Dave Graham is a Technical Consultant with EMC Corporation where he focused on designing/architecting private cloud solutions for commercial customers.