Welcome to the next generation: Symmetrix V-Max is here…

by dave on April 13, 2009


So, it’s 1201am and…well, now you know what we’re talking about, hinting about, laughing about, and honestly…what we’ve been waiting for.  The Symmetrix V-Max is here and it’s ready to drop the hammer on what you know about storage.

What’s in the box?

Simply put, the Symmetrix V-Max ™ is designed from the ground up for scalable growth in virtualized datacenters.  While that sounds more marketing than, well, reality, it isn’t.  When you look at the way the V-Max was designed, you’ll understand and appreciate what the engineering staff had to go through in order to make this happen.  Some highlights:

a.) Porting Enginuity 5774 from PowerPC to Intel x86 code bases (and resulting in Enginuity 5874)

b.) Fundementally restructuring the DMA (direct matrix architecture) into something that was standards-based (around the Rapid I/O fabric protocol stack) while immeasurably scalable (we’re talking hundreds of V-Max engines here…not just 8 like at GA). This is now known as the Virtual Matrix. Every Virtual Engine is tied to the other Virtual Engines via Rapid I/O.

c.) Re-architecting the way that Enginuity interacts with the host OS layers, especially in the virtualization space. Creating the ability to dynamically provision entire port groups, initiator groups, even topologies with less steps than it takes to create a MetaLUN in Navisphere. I’m sure that Chad Sakac will have more to say regarding this but, let’s be REALLY clear on this:  putting a Symmetrix V-Max into your virtualized environment is now going to be even easier than…well, most other things out there.  We’re calling this Auto-provisioning, by the way, and to put it in the words of wizened genius within the walls of EMC (Duane Olson, if I may be so bold):

What a concept.  Create a group for a particular host(ESX farm as an example), and now all you do is create/assign storage to this group and magically, the host has new storage.. One command, and its done.  No more create the device, assign a device to a frontend channel, and then mask the device to a hostOne step, and its all done.  This will significantly decrease the time needed to allocate new storage to an existing host/hosts.

d.) Designing a data tiering system so undeniably radical that the competitive offerings in this space will have to go back to the drawing boards.  <–Ok, so that’s a bit of hyperbole but, hey…give me some credit, I’m excited about this kinda stuff! I’m pretty sure that either Storagezilla or The Storage Anarchist will be getting into the weeds on this.  Suffice it to say, it’s quick, it’s intelligent, and it’s going to give you the kind of flexibility that you need for your data.

e.) Doing away with directors and their fixed methods of integration. Now we’re going to give you the flexibility to add in future SAN topologies while still preserving  your V-Max engine investment.  Oh, and need to scale to more disk? Add an engine!

f.) Extended Synchronous data protection through SRDF/EDP.  This is rather cool.  Rather than have a fully loaded hot side with disk sitting idle, you can simply bang a V-Max engine with cache into place and use it as a SRDF “bridge” to a tertiary site for replication.  From production to secondary, it’s a synchronous replication link and from the secondary to tertiary, it’s asynchronous.  Very cool way to get your “data out of Dodge” as a means of recovery.

Excited yet?  Just wait until you see what we can do with this thing….

Along with the restructuring of the physical makeup of the Symmetrix V-Max, we’re also introducing a new method of doing enterprise storage.  Sure, the Symmetrix high availability, high performance mantra is still in place and honestly, the DMX-4 is still the king of the hill when it comes to enterprise arrays.  What the V-Max brings, however, is 24 x 7 x Forever computing. Think of it as icing on the DMX cake.  It’s a delicious add-on to an already solid foundation.   Rapid I/O links places an emphasis on Virtual Engine to Virtual Engine communication for cache coherency while providing the “glue logic” needed to allow for expansion.  You can start as small as one Virtual Engine and scale, at GA, up to 8 Virtual Engines within a single rack.  You can even provide predictable growth patterns for your storage and add Virtual Engines as needed to manage your scale.  Couple this “growth” process with the ability to use EFDs (enterprise flash drives) in 73Gb, 146Gb, 200Gb, and 400Gb capacities as Tier 0 processing/storage entities along with Fibre Channel and Serial ATA drives for Tier 2 and Tier 3 storage and archival capabilites and, well, you’ve literally got your private cloud in the box.  Couple this together, again, with Cisco‘s UCS systems and VMware‘s vSphere (nee ESX 3.5) and you’ve got a potent data processing solution right at your finger tips. Additionally, the V-Max is designed to scale from a single engine “SE” model to a full 8 Virtual Engines (now) and even more in the future!

Could things get any more exciting?!?!?!

I realize that this is a bit fanatical for me and truthfully, it’s 1018pm the night before and I’m just relishing the suprise, shock, chagrin, and absolutely estatic looks on people’s faces when they realize what we’ve got in store for them and their data.

Are you ready for the next evolution in storage?

Come to the Symmetrix V-Max side….we have cookies! (and Rapid I/O) *grin*



Reblog this post [with Zemanta]