Usually, I don't recap EMC product launches here unless there's an interesting aspect or two that goes beyond the datasheet sort of stuff. After all, EMC announces new things pretty frequently, and I don't want to be one of those blogs that simply regurgitates press releases.
But I thought this one was interesting in two regards: first, the thinking that went into the product, and -- secondly -- a valiant attempt to correct some misperceptions out there.
OK, it's just another small storage array, how interesting could this be?
It depends ...
The product basics
The AX4 heavily leverages the CX3 technology, mostly the FLARE code base. Although it doesn't have the uber-modular upgradeability of the CX3 family, it looks and feels pretty much like a CLARiiON from a technology perspective, and offers a decent amount of upgradeability for its entry price.
BTW, we've been working on making our announcement materials a bit more web-friendly, you can take a look at the results here. Not your average press release, I'd offer. I also like the way the web page looks as well.
What stood out for me with the product were a couple of things: choice of iSCSI or FC connectivity, intermixing SAS and SATA drives to create tiers, a slick out-of-box experience, and very simplified management of some pretty powerful capabilities.
I also like the 60-drive expansion -- nice growth. And when you need a bigger engine, there's the entire CX3 family, which shares the code base with the AX4. "Sharing code base" means that customers get something that's been proven in the market for quite some time, there's a lot of robust features, and -- should you need something bigger at some point -- it'll be very familiar to you.
The primary target
If you're a regular reader, you're probably aware that I tend to stay away from pre-defined marketing categories like "enterprise" or "SMB" or similar. I've found that everyone tends to define these categories a bit differently (so confusion reigns), and -- worse -- customers have this irritating habit of not fitting neatly into marketing categories.
The #1 use case this array is targeting is entry-level production VMware deployments.
Once someone gets beyond tire-kicking and evaluation, they want to build their first serious VMware farm. And, it's pretty obvious that they'll need some sort of networked storage to do this.
Now, there are two schools of thought here.
One is "go low", offer the absolute least expensive device possible to connect up a couple of servers to a shared storage thingie, and off you go. And, of course, in the process, give up some nice-to-have features.
The other approach (which EMC pursued) is to make what we thought the "critical mass" of features needed for this segment as inexpensive as possible. We didn't want to give up on things like high availability, or tiering of storage, or an FC option, or a local/remote replication option.
Why?
We've found that, not surprisingly, smaller VMware deployments take their IT very seriously. They're usually spending for advanced VMware features like DRS and HA and VMmotion. They're usually spending for multiple servers for high availability. And we think that many of them will want a bit more in their storage array than the bare minimums.
If you think about it, consolidation is basically putting all your eggs in one basket, so you tend to look at the basket very carefully. That includes servers and storage.
And we don't think that production VMware deployments are any different -- of any size.
Not everyone agrees with us, naturally
Dave Raffo over at SearchStorage took us to task saying we basically missed the target for SMBs. Dave is entitled to his opinions, of course, but I'd suggest that we weren't going after every SMB on the planet, just an interesting subset with specific requirements.
Dave uses pricing as the primary indicator of SMB-readiness. Yes, costs are important to everyone (including SMBs!) but it's not the only thing, is it?
Dave also seems to think that the previous products (e.g. the AX150) didn't meet expectations. Well, I'm not allowed to refute him with specific numbers, but -- between EMC, Dell and a bunch of our partners -- it did pretty well, thank you.
Dell's got an interesting challenge in positioning the product, since the EqualLogic acquisition still hasn't closed, and they can sell a ton of existing EMC product in the meantime. I'd expect Dell to "announce" their new product lineup (and the positioning) shortly after they close.
And some final thoughts
From where I sit, this looks like a pretty cool array for an interesting segment of customers. Sure, there are lots of choices in this category -- from lots of vendors -- but I think we've got a compelling offering here, and -- all things considered -- I think it'll do very well indeed.
We'll see, won't we?
Product looks great. EMC definitely missed the SMB space - or (apologies Chuck!) the entry level storage space if you wish to stay away from marketing segments - and the critique in my opinion is fair. The reason that other products are doing well in this space - from the mainstream companies HP, IBM and Dell as well as Equallogic, Lefthand, NetApps and the sub 15K vendors (NexSan, Promise, etc.) is primarily due to EMC's lack of offering until now.
Yes, price is important and so is functionality, but some customers invariably just want 'disk', or large capacity, or just acceptable performance RAID and this is a space that EMC has not ever addressed.
Whether this was by lack of understanding of this space or by design for EMC as a company, due to more lucrative targets elsewhere is for you and the rest of the EMC execs to know :)
The AX100/AX150 did not entirely address this space due to lack of scalability and restriction to just SATA drives. The CX3-10 and NS20 are both good products that have gone some of the way to address this space, but are just defeatured or restricted-scalability versions of existing higher-end products in the EMC portfolio.
The introduction of the AX4-5 will partially address this space.
Posted by: mgbrit | January 11, 2008 at 03:42 PM