Looking Back: The 2007 Predictions

I think it's a tradition that if you're a journalist, or even an obscure blogger like me, you do some sort of "predictions" post about this time of year.

Not many people go back and grade themselves a year later and see how good they were at prognosticating the future.

So, rather than offer up a handful of new predictions, I thought I'd go back and see how well I did on the last batch. 

Can't get better unless you know how you're doing, right?

The Starting Point

My original post is from Dec 19, 2006.  I picked a handful of technologies that would probably do well, and a handful that didn't. 

You can read the original here, or you can read the excerpts I'll copy here for convenience.  And, no, I didn't go back and re-edit my post to improve my grades -- that wouldn't be fair, would it?

Pick # 1 -- WAAS

I wrote:

WAAS -- Wide Area Application Services -- this technology tries to help customers get a handle on remote offices that might have a file server or two by consolidating applications, data and information management at a central site. 

At the time, I thought we'd see a lot more WAAS activity in the first part of 2007.  Well, we didn't, not as much as I would have thought.  Sure, there are far many more WAAS-type remote office implementations out there, but I think there's still a gap between opportunity and execution.

Why?  I think it's a hard IT problem for IT to address.  Sure, the technology could always be better, but think about it, if you're going to solve this problem, you're going to spend a lot of money across diverse domains: remote offices, networks, data centers, etc.  And, politically, there will be a lot of voices at the table.

Prediction?  It's going to be a slow roll rather than the more rapid adoption I originally thought. 

Grade?  I'll give myself a B-, I should have known better.

Pick #2 -- Intelligent SAN Replication

I wrote:

Intelligent SAN Replication -- for many years, serious remote replication required a powerful storage array to move information across long distances.  But in 2006, EMC announced a product (RecoverPoint) that uses the power of an intelligent SAN (in this case, the Cisco MDS using the SSM) to perform replication in the network.

I predicted we'd see a lot of people seriously consider this new approach for a variety of reasons, and they'd be happy with what they'd see. 

I was right.  The product has done well in multiple regards, adoption is increasing, and there are many, many successful production sites to go look at today, with many more in the pipeline.

Grade: A

Pick #3 -- Data Dedupe for Backup Part I

I wrote:

Data De-Dupe for Backup Part I -- data de-dupe is understandably the hottest storage technology around these days, simply because it has the potential to do so much for this historical IT challenge.  Not only does it make disk cheaper than tape, it slices backup time, and also has the added benefit of making the information reusable for archiving and repositories.

I was correct here, maybe too correct.  Yes, we saw lots and lots of successful dedupe backup implementations.  At EMC, these were mostly done with Avamar, but at the end of the year, we added dedupe functionality to Networker, creating a "universal backup client" that added yet another trick.

But what I missed was the strong popularity of taking the easy way out -- instead of doing dedupe on the client side, simply plug in a back-end device that does a bit of it at the target, e.g. Data Domain.

I thought that the time savings and network bandwidth saving would strongly favor client-side approaches, such as Avamar.  Yes, it did well, but I underestimated the market for something that wasn't quite as good, but simply dropped in and did its thing.

I should have known better.

Vision grade: A

Execution grade: B-

Pick #3 -- VMware ESX Moves Up

I wrote:

VMware ESX Moves Up -- the vast majority of VMware server implementations I've seen during 2006 have been consolidation of what I call "tier 3" environments: test and dev, ancillary applications, and the like.  VMware's mid-year release of Virtual Infrastructure 3 laid the groundwork for ESX to move up to capture the next tier of application processing in large enterprises.

But new issues crop up that weren't as important before.  Most use cases require a re-thinking of the storage environment, a re-thinking of the management environment, a re-thinking of how you do backup and recovery and so on.  EMC has done a lot of work doing 2006 to engineer its products for use with virtual servers, and I think we'll see that pay off during 2007.

Well, when I wrote this, I was going out on a limb a bit.  Now it looks like a no-brainer, doesn't it?

Not only that, EMC's portfolio is very well positioned to support "VMware at scale" implementations, and we've done quite well there, and the future is very bright.

Grade: A+++

Pick #4 -- eDIscovery Becomes The New ERP

I wrote:

E-Discovery becomes the new ERP -- With the new FRCP rules in place, larger organizations will forced to produce electronic records as part of litigation.  Now, since legal time (and IT time) is expensive, and the consequences can be severe, I think we'll see a boom in strong interest for newer applications that support e-discovery. 

Wrong, wrong, wrong.  I thought FCRP would have teeth, legal departments would wake up, and we'd see intense interest in this amazing capabilities in our regulation-laden, lawsuit-crazy society.

Sure, EMC has done a ton of this during 2007 -- it's a very successful business for us.  But I was thinking of something an order of magnitude larger.  My prediction is turning out to be overstated, both in terms of demand, and timeframe. 

What was I thinking?

Grade: D

Pick #5 -- File Virtualization Becomes De-Facto

I wrote:

File virtualization become de-facto -- as more and more information ends up in file systems, more and more organizations will go looking for solutions that allow them to pool and move resources non-disruptively, and that's file virtualization.

Got this one mostly right.  At the time, I knew what RainFinity could do, I saw the problems many customers were having with file system growth, consolidation and management, I knew how easily it went in, and what the results were.

Does that mean that everyone who needs file virtualization has it?  No, not really.  But most people understand the concept, have figured out if it makes sense for them, see it as evolutionary rather than revolutionary, and are thinking about it in some regard.

And, thankfully, without all the vendor drama associated with storage virtualization -- it's just quietly happening.

Grade: A-

---------------------------------

I also took some time to take some shots at certain technologies that were being hyped at the time that I thought would fall flat during 2008.  It's worth a moment to take a look at those as well.

Anti-Pick #1 -- 8Gb FC SAN

I wrote:

8Gb FC SANs -- yes, the folks who sell you FC components are trying to get people interested in 8Gb FC.  Heck, most customers haven't made the move to 4 Gb yet.  Here's an excellent example of a solution looking for a problem to solve -- at least during 2007.

Enough said on that one, right?

Grade: A

Anti-Pick #2 -- iSCSI SANs in Larger Enterprises

iSCSI in large enterprises -- yes, iSCSI is making strong inroads with smaller arrays and smaller customers, but has hit a brick wall in most bigger shops.  Reason: they've gone with FC, it's installed and it works.  I don't think this will be up for re-evaluation until we start seeing cost-effective 10Gb enet technology that's been optimized for storage, and that most likely wont be during 2007.

A furious industry debate erupted over this one, but -- in the end -- customers decided. If they started with FC SANs, they tend to stick with FC SANs, and are keeping an eye on FCoE.  And VMware adoption isn't changing that, it seems.

Grade: A

Anti-Pick #3 -- Thin Provisioning For Critical Apps

I wrote:

Thin provisioning for critical apps -- as I've discussed before, thin provisioning has its strengths and its weaknesses.  Despite the best efforts of the smaller vendors, I don't think too many people are going to put anything overly important on a thin provisioned environment

I was mostly right on this one.  HDS and NetApp pushed this one very, very hard.  They had some partial successs with less critical apps, but anyone who's anxious about app service levels stayed away from this feature.

Now, that being said, I think there's an opportunity to do thin provisioning better, in such a way that it has all the service level guarantees that critical apps need (predictable performance, better capacity management, integrated with replication etc.) so I'm not sticking with this one in 2008.

Grade: A-

Anti-Pick #4 -- Storage Grids

I wrote:

Storage grids -- despite my best efforts, I've tried to figure out how this particular marketecture might be different than, say, a big clustered NAS environment, or maybe aggressive storage virtualization.  I haven't been able to put my finger on it.  If I can't, I would think that customers won't figure it out either.  Nice buzzword, though -- very sexy.

At the time, there were a few vendors putting up some nice powerpoint on this concept, and it was fun for people to write about.

But absolutely no market traction in 2007.

I think 2008 will be different, though.  Take some the concepts of storage grids, add in SaaS, Web 2.0 and cloud concepts, and it's a different proposition -- watch this one in 2008.

Grade: A

Anti-Pick #5 -- Storage Oriented Encryption and Key Management

I wrote:

Storage-oriented encryption and key management -- most of the solutions available today can only be charitably described as 'tactical', but as we get into the back-end of 2007, we'll see more discussion and the first few implementations of better solutions: wire-speed encryption as part of the switch or HBA, and key management that understands the specific use cases associated with storage.  More here.

I think I was right on this one.  NeoScale folded, NetApp's Decru business stalled, almost nobody bought this stuff.

Why?  Better stuff is coming in 2008 that makes storage encryption an integrated part of the plumbing, rather than an add-on.  Those who waited, smart move, I think.

Grade: A

Anti-Pick #6 -- Data Dedupe For File Systems -- Part I

I wrote:

Data De-Dupe For File Systems Part 1 -- data de-dupe technology also makes sense for file systems that you can read and write normally, but it is more difficult from a technology perspective.  By the 2nd half of 2007, I think you'll see the first few products that expose a file system with de-dupe capabilities: use it for tiering, archiving, as a backup target, test and dev, etc. -- and these will be wildly popular, of course.  I also would expect this to hit its stride in early 2008.

Sure, there's good news, I was right -- NetApp is now offering this product (A-SIS), and it's getting a lot of attention.  But, not so good, EMC doesn't have an offering in this space -- yet.

And it's a big enough opportunity that I think EMC is willing to tradeoff being "first" for being "best".  We'll see.

Vision Grade: A

Execution Grade: C

--------------------

Now, there's more to the original post (go check it out for the other predictions I made), but you get the picture. 

I think I did OK in most areas, but missed a few.  Overall, not bad for an amateur with a second-hand crystal ball.

And, yes, I'm going to take a crack at 2008 predictions, just to see how well I do.

Comments

[tc]

what you think about 2008? i believe its going to be service-oriented-based for storage and backup. The report tools or the architecture of the hardware/software will have to be able to provide utility models, like a subscription-based. Currently, storage is limited by the reporting tools, and backup software is the architecture designed of the software not allowing that.

[Chuck Hollis]

Hi -- I've been giving it some thought, but have also made a point of completely checking out over the holiday season.

I look forward to hopefully writing a post on this in early Jan .. we'll see if I can do as good as I did last year!