All – thank you for all your friendship through 2013. To customers – thank you for being customers! It’s been a crazy year for me (and still lots to finish re: 2014 planning before ringing in the new year), with lots of learning.
I always get a little reflective at the end of the year. I’m blessed to be part of a great team, to be exposed to a lot of customers around the globe, the entire R&D machinery of the EMC federation, and of course all sorts of startup innovation and the EMC M&A operations.
Perhaps that gives my perspective some value, perhaps not… I think that “crystal ball” forecasts are always hokey, but without further ado, here are mine after the break…
And most importantly, have a safe and happy holiday with your friends and family! I resolve to do everything I can in 2014 to make the world a better place in my own little way.
Read on for my top 10 predictions!
Number 10: M&O and “inter domain” skills will continue to climb as the most important ones in infrastructure land.
If I think of the personal Big Hairy Audacious Goals I’m putting in front of myself and my team in 2014, there are two sets of skills that are the most critical. These are: a) “cloud architect” skills – i.e. infrastructure skills that span domains of compute, network, storage,; and b) management and orchestration skills – i.e. the skills to make infrastructure take on dev ops qualities and ultimately turn into IaaS. These two sets of skills are the most precious, rare, and needed at scale. One target I’m putting on the table is to help 4000 customers around the world stand up “well run clouds”. This means self service, elasticity, utility economics, economic transparency – AND using both public and private models (based on workload fit). If I think about the tech – we have the ingredients. We have elastic infrastructure components. We have solid SDDC foundational tech. We have choices of M&O stacks and model based automation tools like Puppet. We have a broad spectrum of public clouds with APIs and SLAs we can all work with. The biggest missing piece are the architects who put it together. At EMC, VMware, Pivotal – we have more than a few, but on all the projects I can think of (and to hit that BHAG of 4000 more by the end of 2014) – the architect/M&O skills are the most precious. If you are one of these people… I. Want. To. Talk. To. You. :-)
Number 9: There will be an insane security backlash, and opportunity.
The whole NSA/Snowden thing has made people stand up en masse and scratch their heads and think about their security and their data, and I get the sense that for many people it’s the first time they’ve thought about it. What do you think is the business model of Facebook, or Twitter? It’s YOU. Furthermore, when you send emails, and make phone calls – of course records are kept at all sorts of places (CDRs for phone info, and of course all those SMTP relays and webmail server logs). I’m not advocating that the government tracking and analyzing all that stuff isn’t something we should be disturbed by – but I can’t compute being “shocked”. Frankly, I think as a human society, having a healthy dialog over who owns information and some business models (and frankly holding the government to their legal limits) would be a very, very healthy thing – and something we’re overdue to discuss.
There are a lot of tools and technologies out there to help – but I have to confess the episode that freaked me out the most was when it came out that the NSA and other entities were working to put backdoors in core crypto algorithms – namely the elliptic curve technique. I suppose that shouldn’t have surprised me either. BTW – some have implied that RSA might have been complicit in those efforts, which RSA categorically denies here.
Between that and putting the first real production quantum computers (the 512-qubit machines made by a Canadian company D-Wave of all places) which could be applied to the “factoring big numbers” problems that are at the root of a lot of crypto – we’re entering a whole new era IMO. The small cynical part of me thinks we are going to find out more and more. The much larger fundamental optimist part of me thinks that we will see a lot of interesting innovation here in 2014 and onwards.
Number 8: EUC will continue a slow evolution and progression. 2014 won’t be the “year of VDI”.
This is almost a humorous topic at this point. Everyone keeps waiting for a “mega moment” around EuC – and it’s just not going to happen like that. Instead, in 2014, we’ll see a continuation of what I saw great progress on in 2013 - “steady eddie” progress on new end user computing models. In 2013 there was some solid progress on remote 3D use cases. I see more and more customers doing real deployments of traditional VDI, and people starting to evaluate new presentation abstraction tech (think Mirage). I see things like XtremIO and other AFA players tackling elements of the challenge. I see VSAN helping at smaller scales (and perhaps more as it GAs, then matures). That’s great progress. It’s not going to be a “big bang”, but rather steady progress and mainstreaming. On a personal level – for my use case, I love my View instance at EMC – but it just can’t replace my traditional client (offline use cases are my forcing function). I think of it as an additional tool (and it is a productivity increaser) – but that makes black and white ROI models harder – and without really, really black and white ROI models, tech deployment tends to follow a slow incremental path.
Number 7: Oracle will continue to stumble, and Microsoft will hit a fork in the road.
I have to confess, I did feel a little vindicated that the Oracle part of my 2013 prediction (here) came to pass when Oracle stumbled mid-year (they did have a good last quarter, so Larry’s still laughing all the way to the bank). There’s simply only so long one can be one of the most reviled vendor by customers (and simultaneously respected as a great traditional RDBMS tech player at the heart of almost every enterprise).
On another note – I just don’t get the “hyper coupling of hardware/software” customer pitch – when I see everything in my world pointing to “abstract, pool, automate” and “decouple every dependency via clean APIs”. I DO get coupled converged (think Vblock) and hyper-converged (think Nutanix, Simplivity, vSphere with VSAN) architectures as the basis for an IaaS layer that is abstracted from the app layer. ..But as soon as you make any infrastructure hyper-coupled to the apps it runs, we’ll I think you’re creating an optimization that has huge downsides.
I’m SURE I have my own sample bias, but it’s also my fundamental world view.
Yet, in the same way that Larry came back to win the Americas Cup – I respect him and Oracle a great deal, and would never underestimate them. And, of course – their database and application stack is strategically relevant to the Business and LOB folks at so many customers, whereas the BEST infrastructure (all the way up to IaaS layers in the stack) are best when they are “invisible”. I just think fundamentally in high tech, when you get arrogant – that’s when you go sideways fast. Particularly where the RDBMS layers are getting increasingly “good enough’ed” with opensource choices at one end, under pressure at the other end by in-memory models. Throw in analytics is moving increasingly from BI/DW to either in-memory at one end and Hadoop at the other end. And… of course, most new apps (enterprise and otherwise) are built on PaaS and frameworks that have less and less to do with “traditional app stacks”. There are a lot of smart people at Oracle – the question to me is whether they can self-disrupt and regain the love of their customers.
Conversely, my prediction on Microsoft was a total miss (which perhaps highlights the value of all these predictions :-) I bought myself and the family two new Surface Pro 2 devices (in the vein of “give them a shot” and “kinda rooting for the underdog”). You CAN see them trying to become the consumer device company, and if you squint hard enough – you can see the promise. But as an end-user – just an individual like others – it’s been an epic #FAIL, for all the reasons people have discussed online. It’s filled with visible obvious “sacrifices'/compromises” – less on the HW (which is solid)and more on Windows 8.1. I’m going to keep at it, and keep giving them a shot. Surface team – if you want to talk, I’m here :-) I’m happy with my Xbox One – and I think Microsoft has another winner on their hands if they can keep content flowing and the new Kinect sensor is really quite awesome. Azure is really starting to crank. I think the challenge they are having is the lack of focus and intensity. Personally I think that MIGHT be an argument for splitting the company, or doing a “federation” like thing of their own. A huge factor in the world of IT will be their CEO pick. They are really at a fork in the road. I’m rooting for you Microsoft – prove me right, and disrupt yourselves. Roll big.
Number 6: SDS will disrupt the storage industry (even more than Flash). SDN will start to become more mainstream.
While there is a ton of disruption in storage land, the big impact is not from the AFA and flash media transition (though that is huge), it’s around new architectural models. By that I mean software only data planes – across use cases of transactional hot data (think ScaleIO and in 100% VMware use cases VSAN), software-only NAS, software only object stores that are behind the huge content depots and new applications. Also – storage abstraction/pooling/policy is going to be a big deal – as without this, the variation in data planes drives complexity and rigity. People are internalizing no “single dataplane” is right because (the more workloads a storage system is well suited for) = (the more “ ediocre” it is for any given workload).
That doesn’t mean “platforms that do lots moderately well” is bad – that’s the huge, huge market VNX, NetApp and others serve. Another way to think of it is this: so long as you are ok (or have less workload variation), then you can reduce down your number of storage architectures – and at one extreme, in some cases can do it all with a single very general purpose platform.
… But there will be NO storage architecture that deals with all models of data persistence in an ideal way. I want everyone to have perspective here. Today the world of “external hybrid storage” is about at $60-70B market, and continues to grow. BTW when it comes to hybrids including EMC with Isilon, VNX, and VMAX – many can come in “all flash” variations. Some do large amounts of flash extremely well (better than others).
Calling a purpose built AFA like some do strikes me as wrong – as it does others. I tend to agree with Chris’ point – that an array that just has a ton (maybe even 100%) flash doesn’t make it architected for flash. Small changes don’t cut it in my opinion either. It doesn’t mean it’s bad to start “optimizing” traditional architectures for larger and larger flash mix (a next gen VNX with all flash can deliver high IOps and low latencies, even if it’s not architected for all-flash – analagous to the 3PAR point Chris was making)…
But back to the point – ok $60-70B of hybrids, including mega content depots = a huge market. So – what’s the size of the “purpose built AFA” market? By the end of next year, if you add up all the AFA players (including EMC with XtremIO) will be a $1B – MAYBE a $2B market. How many startups can that market sustain, even at high CAGRs? (see my later point on “AFA armageddon”). The “SDS” market will be much, much smaller.
So – if the pure SDS market is smaller than the AFA market, why do I think that SDS models are even more disruptive? Because they disrupt the core market – not just the tech. They have a different economic model (high margin software at lower top-line revenues, with a “bring your own COTS hardware” element). They mean that new architectures (hyper-converged) are possible. They mean that people who are big players – but not storage vendors – can try their hand in the game.
This will affect not only the big folks – but also the smaller new players. While I’m sure Nutanix and Simplivity would prefer the hardware appliance model – I would suspect the pressure to do “software only” variation is high. This may seem as all good – but as a startup it changes the model a lot (different economics – lower top line revenue, higher margin models). Over time – SDS (control plane and data plane) will have big echoing changes across the market – and you’ll see more and more of that in 2014.
SDN battles are already being waged furiously – and enough ink has been spent on that that I don’t think I’ll add much. One thing that I will say – I think customers should use 2014 to learn, play and use these SDN/SDS models. Part of their value and appeal is – they are SOFTWARE people! Anyone saying “buy my hardware it’s software defined” needs a smack. I **DO** think that it’s legit to point out hardware that is designed to “work better” in an software defined policy world – but we’ll see how that plays out. Play, learn, try – which to a technologist is music. I’m pushing like mad inside EMC to get all of our software-only assets out there in the “software company model” – simple and easy to get access to. Long way to go – but if we do – I think there’s some incredible things we could do!
Number 5: Big Data and Fast Data will continue their inexorable impact/disruption wave.
In 2012, we saw the first national election that you could say was “won in large part through big data and analytics”. There’s fascinating tech here. In the land of Big Data, it’s mostly in the open source domain, with innovation happening at Pivotal, Cloudera, Hortonworks, Intel and more. Likewise in Fast Data (in memory data models) – super cool stuff in both the opensource and commercial worlds. This links to my earlier point where I see these coming up more and more in Enterprises, and putting the “ACID RDBMS as a universal source of truth” under compression. The key here isn’t tech, but people. Like “Number 10” on this list – skills here are valuable. Within EMC as a whole, Pivotal is running furiously on this topic – and the EMC Information Infrastructure part of the federation is spending a TON of R&D into object and HDFS software-only stacks.
Number 4: People will continue to put “Public cloud” FUD into “that’s stupid” category and move forward – but also learn more and more about what’s a fit, and what isn’t.
This one is simple. More and more people are realizing that “Public Cloud” is secure. More and more “industry specific” public clouds are popping up to deal with specific industry constraints and policies. More and more people are realizing the public cloud options have great economic upside do to elasticity. BUT the flipside is also true. People are realizing that elasticity has a huge pricetag. When you have a steady state workload, or workloads that are not easily made software-resilient (aka legacy enterprise apps), building a well-run private cloud is better AND cheaper.
Don’t get me wrong (see Number 2) – Public Cloud and SaaS are absolutely putting pressure on traditional on premise IT. That’s not my point – it’s that we will see customers use both private and Public/SaaS models liberally.
This realization is creeping into the broader “zeitgeist” of IT land – and I’m glad. I’m sick and tired of silly “either/or” arguments on this topic. Outside people with an angle, people with twisted perspectives, or media folks who don’t know better and are looking for a headline – people know this is an “and”, and IT will be a broker of both :-)
Number 3: “AFA Armageddon” will continue.
Violin is now positively imploding (see here and here). FusionIO likewise is in trouble (if not collapsing to the same degree as Violin). Both of these companies where stock market darlings a short time ago. What’s the scoop? The trick of doing things to rapidly gain mind/market/wallet share is that it becomes difficult to reverse course. When you establish a certain burn rate (cost of operating a certain way) and business model – the “I’ll make it up in volume” can actually be true – IF you get some way to grow your revenue at a lower and lower cost model. When Violin lost their HP OEM (Volume with a low cost of sales), their already fragile business model collapsed. FusionIO came under pressure as the cloud-scale customers (small number of customers buying a lot of stuff = low cost of sales) started to defer or look to lower-cost commodity hardware players.
There’s something to be learnt here. If a startup is burning money furiously, and operating at a considerable loss (this is normal in early stage startups) – the question is: “what conditions will change that will dramatically improve your economic model?”
I know that this sounds pessimistic – it’s not. Flash is a disruptive core tech. There are billions at play – so it’s easy to get funded. But – the business is brutal. The “death rate” of the flash startups will be quite severe. Even the most successful of some of the startups (see Nimble – who just IPOed – see here – but is still losing $30M annually on $89M in revenues) suffer from this condition. Growing revenue is good. Doing it while loosing more and more money proportionally is not so good.
Also – I would expect this area to become more competitive over time. With EMC in the game with XtremIO (which I think is defensibly one of the stronger solutions in the space – though it of course has room to improve, and you can count on a rich roadmap through 2014), with others expected to join in 2014 – I don’t see how the AFA market will get easier. Does this mean that there can’t be successful new startups in the storage domain? Of course not (and frankly, I love startups). Just expect “AFA Armageddon” to continue.
Number 2: Industry “reconfiguration” will continue – and accelerate.
I’m an IDC fan. They do a lot of analysis, and quantitative analysis (not “hey what’s our opinion”). They published a paper called “IDC Predictions 2014: Battles for Dominance — and Survival — on the 3rd Platform”. So – this isn’t a Chad prediction – rather an IDC one. That said – I’ve been furiously working on this within EMC, and this quote struck me as so apropos, and so concise, I want to reference it (and I would encourage you to buy the report – it’s only $5 and worth it):
In 2014, the companies that won market leadership in the 2nd Platform era — and are the incumbents in most customer sites today — will even more aggressively reconfigure themselves for the fight for the 3rd Platform marketplace: acquiring more 3rd Platform competencies and divesting diminishing return businesses (or businesses that are marginal to the new core). Companies likely to be very active acquirers and divestors in 2014 include Microsoft, IBM, HP, Dell, EMC, Cisco, Oracle, SAP, AT&T, and Verizon.”
I can say with confidence that we will use every resource at our fingertips, every ounce of innovation we can muster, our unique Federation business model to do exactly that. I expect no less from other IT leaders.
Number 1: We will see amazing things that no one expects – including me!
This is always my favorite part of ringing in the new year (except seeing my family!)… Whenever I toast au revoir to the last year, and bonjour to the new one – I think to myself of all the great surprises, learning, growth that were UNEXPECTED in the previous year. In many cases, things that looked like grey thunderclouds had silver linings. 2014 will be no different, and I encourage us all to embrace change, embrace suprise, embrace adversity, embrace each other – and it will be the best year on record!
Happy Holidays, Merry Christmas, Happy New Year!
Chad Sakac, Dec 23rd, 2014