« VMAX and vSphere VAAI XCOPY update! | Main | Goodbye 2012, Welcome 2013Happy New Year! »

December 21, 2012


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

Matt COwger

Speaking of apple - I do believe I won our bet about Apple's stock price 1 year after Jobs no longer leads the company :)

Greg Knieriemen

Great predictions.

One thing I wonder about (and this isn't a loaded questions): Who will be the first with a complete Big Data (end-to-end, converged?)solution and will that happen in 2013?

Henry PAN

Let's see how many fo those prodection will come ture in 2013:>)

Merry Christmas

Stu Fox

Hmmm, correct me if I'm wrong but vSphere replication & Hyper-V replica are VERY similar. In fact I'd almost argue that Hyper-V replica is better since it doesn't need a virtual appliance, it is truly hypervisor integrated. VMware have better configurability of the replication period (Hyper-V is set at 5 minutes) but interestingly it's only every 15 minutes minimum unless I read the documentation wrong. Interesting that VMware's site say that "it is the only truly “hypervisor level” replication engine available today". Probably another case of VMware trying to get the market to believe something just by saying it (they're good at it too).
Microsoft have tried the "works better on Hyper-V" thing, but they've never managed to get it together to actually prove that. And I guess that's hard, VMware are unlikely to approve benchmarks run by Microsoft showing their workloads run better, so even if it was true they still couldn't prove it.

Be nice if they can get Storage DRS like in place for the next version - the underlying technology is there (live storage migrations plus storage offload).

iwan rahabok

I love #8, especially this part: "For infrastructure – the closer you are to invisible, the better."
I see 3 stacks: Service, Apps, Infra.
Infra is the bottom stack. For us infra team to be more valuable to the business, we need to turn the Infra into a Service.

Have a great 2013!


Chad, only you could work "crazy Ivan" into a tech discussion and have it make perfect sense! I think most of your predictions are spot on. I also think that 2013 is the year that Enterprises move towards SDDC and Cloud in a big way. I also predict that 90% of the players in the market will NOT figure out that this is 10% technology and 90% people/process.


Just as a quick correction - Hyper-V does have replication just like vSphere. I haven't personally tested it's scalability, so you could argue it's not as refined, but it's definitely there. And while they don't have SRM, you can get very similar functionality with some simple powershell scripting. Example:

I'm also confused about your Openstack comments, the Folsom release added support for Hyper-V 2012. And from a networking perspective, most of the networking guys I know think Hyper-V's implementation of WNV/NVGRE is actually superior to VXLan... so can you expand on how you think they're being left behind? Case in point:

Travis F

The list of traits is good and warrants a follow-up. All the stars reading you blog want to know how to get recognized and how to recognize other birght lights.

Knowing how to spot the traits is equally important and valuable. Do you have tips for recognizing and evaluating traits? I suspect it's challenging to everyone who takes it seriously.


Nice predictions. In my humble opinion, these are the trends, traits and skills that will herald a new paradigm in 2013 (more so in traditional enterprises as opposed to startup-types who are already on the spearhead of that paradigm) --

Technology trend -- Distributed computing, more de-centralized model of compute (discrete processing, storage) tied together via software

Skills -- Devops (where automation, optimized performance etc will be developed and managed by the same team)

Traits -- Ability to look beyond the status quo (you may call it "dreamer", I call it thought-leaders)

Happy New Year. BTW, I've been a silent reader of your blog for at least a couple of years now. Great stuff...helps understand general trends in the industry!


I contend that both GlusterFS and Ceph (and maybe a few others) do qualify as software defined storage. Policy like "do I replicate this object" is indeed decoupled from the hardware in these systems, and works across vendors, so I think they fit your definition. (Disclaimer: I'm a GlusterFS developer.) When the phrase "software defined storage" first occurred to me, I thought it seemed like a really cool hook for us to use. Then I saw that others were picking it up and, in a rare moment of hipsterism, abandoned it as hackneyed before it even became popular.

I also slightly disagree that your assertion that people will rush to leverage HDFS. If they do, they'll run up against its limitations and run away again just as fast. Many, including EMC's own Greenplum[1], have articulated a different vision of federating storage that already exists instead of migrating onto storage that's unsuited to any task but one, and I think they have it right[2].

[1] http://www.greenplum.com/blog/topics/hadoop/hadoop-and-disparate-data-stores
[2] http://pl.atyp.us/wordpress/index.php/2012/11/the-future-of-storage-for-big-data/


Apple are not going to rollover and die. They are a 450 billion dollar company, who will re-innovate starting by acquisitions in 2013.

They may come knocking on EMC and VMware's door to assist you in keeping Microsoft and HP down.

Money talks!

Paul Whyton

Great musings as always and i personally feel the SDDC will become the primary focus at least of discussions if not implementation. A lot will depend on the emergence of automation tools that are vendor agnostic as more customers will have a heterogenous Datacentre either through choice/design or organic growth.
Much will depend on the underlying networking design and coupled with Security I feel these will be the 2 focus areas for those of us in a consulting role.

Chad Sakac

@Matt - indeed, you won that wager - what were your terms again?

@Greg - I don't think ANYONE will have a complete Big Data converged solution in 2013 (or '14 for that matter). what's going on in the big data space is still in that early "explosion" phase - where a ton of value is being generated in funky unique ways. I would expect that in '14, we'll start to see consolidation of IP and companies to provide more "integrated" value.

@Stu Fox - you are right, I'm wrong - they DO have integrated replication in Hyper-V. Thanks, and I will fix the post. That said, that wasn't the main crux of the point (in fact, it might strengthen my point) - Hyper-V is a FINE product, and the stumbles are simply one of Microsoft not rallying around it. Imagine how much they could drive by shipping Exchange, SQL Server, Sharepoint and more by **default** as virtual appliances, and optimizing for Hyper-V. They could do this without acting in an anti-trust fashion (as it doesn't exclude those apps running on vSphere or Xen), but simply making them work BEST on Hyper-V. Yet, strangely (and I think this is the "business unit" structure of Microsoft showing up) - each of the app teams almost seems to consciously **ignore** hyper-V's existence.

@Iwan - thanks for the comment! I completely agree!

@vDirtyBird - you rock, my friend! Happy 2013 to us, and I love working with you bro!

@TimC - thanks for the comment! re: Hyper-V replication - thanks for the catch, and see my comments to @Stu Fox above. Re Networking - I don't think I ever said that Hyper-V and Openstack were falling behind? My openstack comment was that I'm seeing them more and more (mostly in the service provider and education markets, less within Enterprise). Further - I predict in the post that we'll see more and more folks (certainly VMware and EMC) embrace Openstack. My Hyper-V comments were that it's success seems limited by Microsoft's own embrace of Hyper-V in their GTM. From a technology standpoint, it's certainly a viable option for many.

@Travis F - thanks for the comment! You know, I haven't codified in my own mind how to evaluate and recognize those traits. For now, the principle I try to use is a simple one (but seems to work): a) manifest those traits as much as I can as a leader; b) select employees personally based on those traits; c) emphasize them to model those traits as much as possible - I hope that A+B+C = it echoes through the broader organization.

@Implace_order - thanks for the comment, and the readership!

@Obdurodon - thanks for the comment. And, btw - thanks for your development work on GlusterFS - it's pretty cool! I disagree (I think) with the point you're making - and I think that it's a battle that will play out in 2013 a lot. I **don't** think it's sufficient for a storage target to run as software on COTS to be called "software defined storage". That principle, applied - would mean that GlusterFS and Ceph - while very cool, and software that runs on COTS hardware wouldn't be "SDS". Just to make it clear that I'm not picking on anyone in particular - I would say the same thing if EMC put out Isilon as pure software (bring your own CPU, network, persistent storage), Atmos, VMAX, or VNX.

I **DO** think that we can expect to see more and more storage stacks as software + bring your own COTS in the market - I'm not poo-poohing it. As an EMCer - I'm thankful that we started a long (and not easy) shift to embrace commodity hardware components and pure x86 software stacks - this makes our IP more able to be leveraged in different ways in the coming years.

That said - I **do** think that there is a couple defining elements of "true" (seriously - heavy emphasis on the quotes there) of "SDS". I think the principle that is needed desperately isn't just the decoupling of the stack value from the hardware (the point I think you're arguing, which I whole hearted agree with), but also that there is some standardization of the "control plane" API for storage - the "OpenFlow" equivalent.

Now - the point of "can there be value in the hardware" - I think the answer of that will always be "yes, to some extent". I'll give you an example that you might not think of. Within an enterprise (and service providers are different), hardware mech replacement, local sparing, density - those are all things where people/business do various things - and so long as customers see value in it, fine. Is it where they bulk of the value is? Nope.

@vmcreator - I **CERTAINLY** didn't say they would roll over and die. I said that I think they will "stumble". When you are a public company - stumbles are surprisingly easy (and tend to get blown up in the media and your stock price). They will re-innovate and re-invent as they have before and will again.

@Paul - thanks for the comment!

Wade O'Harrow

Did I really just read what you said about MicroSoft and Hardware / Software ??

Great read.. Nice to see that somewhere in there you are still yourself. We did learn a lot over the past 5 years didn't we?

Keep being the Leader, the minute you stop stepping out on the edge and fearlessly looking across the abyss with un-waivering optimism, you will become a manager and cease being a leader. This would be a very sad day, please don't let it happen...!! "Cynics and Skeptics, withered and ran from his gaze....."

Brian Verkley

From the audio version, you said, “Big Data is going to change the world as much as cloud did.”
What do you think will drive more change, a technology that allows a company to reduce cost or a technology that allows a company to create new revenue? In my mind, new revenue. Every time.

And hey, I was a web developer in 1995. No time travel required :-)

The comments to this entry are closed.

  • BlogWithIntegrity.com


  • The opinions expressed here are my personal opinions. Content published here is not read or approved in advance by Dell Technologies and does not necessarily reflect the views and opinions of Dell Technologies or any part of Dell Technologies. This is my blog, it is not an Dell Technologies blog.