Peering into 2013 and Beyond

By on 12/21/2012.

nostradamus.lge  Peering into 2013 and Beyond

As has become [almost] the status quo for tech bloggers, it’s borderline-mandatory that we do a “predictions for the next year” type of post, so that we can look back on it and laugh the following year.  I didn’t get around to doing one for 2012, so I want to make sure I get one in for 2013.   Folks, from what I know that a lot of you are not privvy to, this is going to be one helluva ride!

DISCLAIMER:  The views and opinions I’m going to express here are mine, Nick Howell, of DatacenterDude.com.  While I am an employee of NetApp, these are not reviewed and are completely unrelated to any product announcements or company directives.  These are simply my personal opinions of the industry at-large.

1) Ethernet and file-based storage will rule the day.

I’m not just talking about virtualized environments, either.  If you’re still one of those shops heavily relying on “fake 16Gb” fibre-channel because you’re still under the oh-too-common misconception that 10GbE NFS is somehow less performant, buddy…you don’t know what you’re missing.   Throw away those SCSI handcuffs, forget about managing LUNs, zoning switches, and all the other mgmt headaches that go hand-in-hand with FC, and drink this Kool-aid.  You’ll thank me.  At the same time, do it smartly.  What do I mean by that?  You’ll hear a lot of companies telling you that block is still the way to go, but that’s mostly because they require some sort of gateway device or that they really don’t have an easy solution for file-based protocols like SMB/NFS.   Make sure you’re doing your homework, reading between the lines, and REALLY looking into what you’re getting when a vendor is promising you the world.  I mean that for all companies, NetApp included.

Bottom-line:  Look at some data from 2000-present.  Fibre-Channel adoption has leveled off, if not declined, while Ethernet has exploded since around 2008.  Guess what happened in 2008?  vSphere 4 with NFS support, and Cisco NEXUS switches started taking hold, with 10GbE capability.   Alcatel is already doing long-haul 100GbE line cards, and Cisco is using Palo cards in UCS for 40GbE logical pipes.  Won’t be long before we have 100GbE PCIx (or whatever standard is out then) cards to put in the servers and storage.

2) VMware has peaked.

And I don’t mean this in a derogatory way.  They’ve done just about everything you can, including trying to eat in on some of their partners’ space (vSAN, Nicira, et al), to virtualize a datacenter.  vCloud Suite 5.1 and beyond is the culmination of all of that.  This trend we’re seeing of them focusing on Automation (vCAC/vCO) and End-User Computing (Horizon/HorizonDATA), to me, is more of a lateral, horizontal move into consumerization of the enterprise.  And I think they might not be wrong here.

If you’ll allow me, I’ll explain.  We’re getting to the point where hardware, across-the-board, is a commodity.  It’s infrastructure.  Silicon and flashing lights in a metal case, with the sole purpose of processing 1′s and 0′s.  Which means absolutely JACK-freakin’-SQUAT to Betty Sue managing her spreadsheets on the Sharepoint server that is currently down because a storage failover didn’t happen, or a NIC popped, and HA contingences weren’t implemented properly.

What’s my point?  By focusing on the end user side of the house, [hopefully] VMware is getting CIO’s and Director’s to re-focus on the end user side of the house, rather than a whole team of folks managing infrastructure (you know the stereotype…the trolls in the back closet that manage the servers/storage and never come out).  Deliver applications.  Automate daily infrastructure and helpdesk tasks, such as deployment of servers/storage/endpoints.

We’re allllmost to that point where everybody has jumped on the private cloud wagon, and customers have P2V’ed most stuff in the datacenters.   Believe it or not, there are plenty out there who haven’t, I talk to some at least once a month that are still developing a “virtualization strategy,” both big and small, but that always going to be the case.   We’ve got to start thinking smarter about how we deliver content and accesibility to end users.  Period.  Apple changed the world with Xcode, Apps, and the App Store.   I see VMware really trying to do the same thing here.  Almost making the exact same play.  But beware, you’re going to run into the same three problems I always warn you about:  People, Process, and Politics.  None of which have anything to do with infrastructure, nor application delivery.

3) Holy Microsoft deja vu in 2013!  Only this time, it’s VMware getting overrun, not Apple.

They’re on a tear right now.  Your tablets, Windows Phones, and Windows desktops now look like your Xbox dashboard.  SMB 3, Hyper-V 3, and Win2012 are going to make a giant, booming, Mayan-apocalyptic splash, and both Apple and VMware are going to feel it, in both consumer and enterprise markets.  With VMware shifting focus to end user computing and consumerization of the enterprise (my projection), Hyper-V will begin to takeover some SERIOUS market-share in the infrastructure space that VMware has dominated for the long-term.  They are already showing some 70%+ growth numbers, and I think that’s only going to continue to chip away.

Apple, on a side note… I have a personal bone to pick with you.  Your $3,000+ devices are not “disposable,” and you’re pissing a lot of us off by soldering everything to the motherboard.   STOP IT!  It might already be too late.  You’re turning the Macbook Pro and iMac line into glorified giant iPad’s.  And unless you’re going to start selling them for < $1,000 I’m going to start entertaining running Windows again because I can modularly upgrade the hard drives and memory.   And that has a trickle down effect.  ‘Cause them I’m gonna want to go back to using my Windows Media Center setup instead of my iTunes + AppleTV, which means I won’t need that Mac Mini I’ve got anymore.

See where I’m going?  Some of us out here need to interchange these components over the lifecycle of a machine.  I don’t need it paper thin, and light as a feather if it’s going to be outdated in a year.

4) Clustered ONTAP paves the way to a whole new level of scaled infrastructure.

Look, DataONTAP is already the #1 deployed storage operating system by a large margin.  It’s no long stretch that within 2-3 years this will be the defacto standard of how storage is deployed, setting a model for others to follow.   Make no mistake, NetApp IS innovating, contrary to what you may be hearing from the competitors out in the marketplace.  Trust me, they’re not telling you anymore than you already know.  There’s reasons for them not knowing what we’re up to behind closed doors.  And for those reasons I cannot share with you anything other than ……..

5) What if storage controllers could also be, or even house, server hypervisors?

I’ve had this theory for some time now.  It’s just software running on processors and memory, right?  Is it too far of a stretch to see some type of convergence of host-based hypervisors onto large-scale storage controller arrays/clusters?  No, not really.  For the same reasons Virtual Storage Appliances are making a big splash currently.  One of the hottest topics at our Insight events this years was DataONTAP Edge (our VMware-based VSA of DataONTAP).  What’s to say we couldn’t do the opposite?  I think we’re going to see some of this in the near future.  And no, I’m not talking about blade chassis’ here, but similar.

6)  OpenStack will have it’s 15 minutes, but that’s about it.

If you’ve logged into twitter or read any number of tech blogs recently, you’d be hard-pressed not to have heard some mention of OpenStack.  Personally, grassroots growth and community-driven passion to do things their own way makes my hair stand up.  I love that stuff.  But as we’ve seen over the decades now, we’ve watched what Richard Stallman & Co has done with the Free Software Foundation (FSF), we’ve watched Linus Torvalds and the Linux community grow into a behemoth, but somehow not really take hold, all while Microsoft, Apple, Sun/Oracle, and VMware have absolutely DOMINATED the market in the last 20 years.

I believe OpenStack will have the same experience.  All the major vendors are saying, “Oh yea!  We’re partnering/playing too!” but that’s mostly on a contingency (in my opinion) that it does take off.  Personally, I don’t think it will.  And frankly, I hope I’m wrong on this one.

7) The triumphant (or woeful) return of the Storage TwitPiss.

2013 is going to be a big year for storage.  Clustered ONTAP adoption is ramping up.  I’m looking at current numbers and seeing many deployments already, some of which I know for a fact are production (see: our recent story with CERN).  In the past, twitter was a battleground of storage companies, namely EMC & NetApp vieing for the “street cred” of the internet as to who was better.   Given the level of innovations both companies are doing, I wouldn’t be surprised if we got quite a few of the “NUH UH! OURS IS BETTER!” type of fights, better know as “TwitPiss’es” on Twitter.

I loathe this type of stuff, because it does the customer no good at the end of the day.  I want to educate, evangelize, and share passion about products.  But unfortunately, I feel within the force that we’ll be back at it all over again.  Ugh.

8) Tape continues to help people sleep better at night, but that’s about it.

Are you still using tape?  C’mon, you can tell us.  If you are, I’d like to chew on your ear for half an hour or so, as an ex-admin/customer that found a way to do it better, faster, more efficiently, and with a self-service model.

Will tape still be around?  Yes.  But only for the same reason’s it’s continued to be around for the past 4 or 5 years.  It’s one of those things Director’s and CIO’s think they have to have.  Step up, admins/engineers, and let’s show your boss how we can revolutionize your backup strategy.  It was the NUMBER ONE PRIORITY for 2012 from an IDC CIO survey, to Improve Backup & Recovery.  Where does your boss stand?  Where do YOU stand?  Let’s get you out of the business of managing a 24-hour non-stop tape rotation, save you a ton of money on tapes and Iron Mountain pickups/storage lockers, and improve the efficiency of your organization.

It’s a lot easier than you think.

Those are the big one’s from me.  How about you guys?  Would love to hear your thoughts on my predictions, or even hear you make some of your own!

-Nick

8 Comments

  1. If Cluster DataOntap is the future I would like to see NetApp have a path from 7-mode to Cluster Mode . What will happen to 7-mode ? We just made an investment in Controllers running in 7-mode .
     

    • I agree that there does need to be a cleaner migration path from one to the other.  The good news is that it’s just software and can be run on the same gear that you’ve likely just made an investment in.  I cannot speak to the future of 7-mode.  What I meant from the projection is that Clustered ONTAP will pave the way by leading the industry down a path of true scale-out, software-defined, cloud storage.  Much the same way VMware did for server virtualization.

  2. The Tale of the Tape

  3. Good post Nick – I appreciate how much effort it takes to coalesce your thoughts on varying and complex topics into a succinct blogpost! I agree with much of what you’ve said – ethernet will increasingly rule the day but it’s going to be quite a few years before FC and SCSI based transports go away.

    Have VMware peaked? The software defined datacenter needs to happen if cloud computing is to fulfill it’s promise but as others have pointed out (http://blog.fosketts.net/2012/11/29/follow-yellow-brick-road-softwaredefined-future/) there’s a clash of business models -  EMC are the world’s largest storage company by valuation and a quick transition to new architectures and solutions isn’t in their interests (or Netapp’s for that matter). I’m probably too close to VMware technologies to be completely objective
    but the million dollar question is whether VMware can capitalise on that shift or whether it’ll be the existing incumbents or new startups that capture mindshare and marketshare.

    Storage controllers converged with hypervisors is an interesting theory and given the popularity of converged infrastructure quite possible. Will it suit all customers? Probably not, but it’ll be another tool in the kit bag to be used where appropriate. At the moment I see a shift towards distributed architectures more than centralised – if you’re investing hundreds of thousands of dollars on big iron do you really want it to be all things to all men?

    I’m more open minded about OpenStack. Open source is getting increasing traction as the guest OS becomes less relevant but unlike the financial returns from Apple, Microsoft, VMware etc there’s no single measure of it’s success. Many of the Web 2.0 generation companies like Amazon, Facebook, Netflix etc have built huge companies from open source building blocks (or generated open source in the process) and given cloud computing’s benefit from scale I think OpenStack could succeed. It’ll take a few years to mature but don’t count it out yet.

    As for storage arguments maybe there’s a silver lining – it’s a sign of a lively industry when there’s healthy discussion!

    • @490c21286c1e9c7639d9d03bddb5e143:disqus ,

      Thanks for the comments!  One of the things I forgot to talk about was “cloud” and what was “cloudy” to me and what is not.  I think the usage gets blurred too often.   To me, there is “converged infrastructure,” which will often times get labeled as a “private cloud,” and then there is AWS and Google (as well as some other players) doing the public cloud stuff.

      Enterprise lives in the converged infrastructure/private cloud space.  I don’t think this is going away tomorrow, but it’s not going to be around for much longer, relatively speaking.  Enterprises, and traditional IT vendors, are going to have to find [yet again] new ways to do IT. See our partnership with Amazon if you want a glimpse of the future.   Us vendors are not “in trouble” as people like to tout.  All that stuff still has to get stored somewhere, and whether it’s getting stored and replicated in YOUR datacenters, or in the “cloud’s” datacenters, the point is, it’s still getting stored and replicated.

      I bet this is super-frustrating for CIO’s and Directors the more I think about it…

      “Yea but they told us to virtualize everything and we would be ok!  We did and now we have to change again?!”

      “Don’t we already have our own cloud?! No?  What we have isn’t technically a ‘cloud’?”

      “UGH.”

      It’s no secret that I’m one of Netflix’s biggest fans.  My absolute, hands-down, fav blog to read is techblog.netflix.com.   Go over there and spend an hour, and you can quite literally WATCH “cloud” getting invented.  Half the stuff those guys are doing I haven’t even heard of yet, and they’re publishing it all to a very structured github community.  So you’re right.  Open Source is certainly enjoying a moment in the limelight, and I feel it’s only the beginning.

      Where I stand with OpenStack specifically more revolves around the sheer number of distros you’re going to end up with, saturating, splitting-off, and blurring the technicolor.  I love open source work, but at the same time, it is its own worst enemy by not consolidating efforts into a single stream of consciousness.  I will say that I do like the approach of “we are the kernel” pluggable framework that its becoming.  I’d like to see some clearer messaging around that.  But so far I’ve heard of at least 5 or more companies that were all “building distros.”

      Check out The Cloudcast podcast with Aaron Delp and Brian Gracely.  They got to sit down with one of the original creators from Nebula, and he tells the WHOLE story of how OpenStack came to be.

  4. Datacenter Dude - Predictions and The Bowl Championship Series - Datacenter Dude

  5. #28 2013 Year In Review - Datacenter Dude

  6. I’m amazed, I have to admit. Seldom do I encounter a blog
    that’s both equally educative and interesting, and let me tell you, you’ve
    hit the nail on the head. The issue is something which too
    few folks are speaking intelligently about. I’m very happy that I came
    across this in my search for something concerning this.

Leave a Comment

Your email address will not be published. Required fields are marked with *.