Thanks to all of you who have been following the EMC World 2011 coverage -- from all of us
bloggers! I know, there's a lot to read and contemplate, but things are moving very fast indeed ...
I thought I'd use this post to recap the core ideas shared at the outset of the event. Most of this material found its way into Joe Tucci's keynote, but you'll recognize a few slides from other keynotes as well.
For those of you who are avid followers of this blog, many of the themes and concepts should be familiar. But here, they're all in one place.
Simply put: if you'd like to understand how EMC views the world going forward at a macro level, this is the blog post you want to read.
To Begin With
Our foundational belief is that the IT industry is undergoing multiple transitions at the exact same time. For those of us who've been around for a while, there is no precedent we can easily point to.
Part of this was represented in the theme of the show: "Cloud Meets Big Data".
But there's far more to it than that ...
"Cloud", in this context, represents a convenient shorthand on how we see IT services being delivered in the future. And "big data" represents the tantalizing new source of value we're all seeing.
Joe started his keynote by comparing the theme of last year's EMC World ("Journey To The Private Cloud") with the motivation this year: accelerating your journey to the cloud.
Last year, people were larging evaluating the concepts and ideas; this year it's all about execution.
Joe was also very clear in sharing how he saw EMC's role in this: "... to lead customers on their journey to cloud computing and transforming IT".
Put differently, this is less about selling spiffy new technology, and more about helping our customers and partners through an important industry transition.
Refreshing.
Cloud Motivations
Joe then restated familiar ground around reasons why so many IT organizations are moving in a cloud direction.
First -- and most visible -- was the lopsided IT spending profile. You've probably seen this slide -- or something similar -- many times before.
The message is simple: far too much gets spent on maintaining steady-state IT functions, and far too little gets spent on creating new value from IT. Whether you decide to spend more, less or the same amount on IT is a separate decision: but just about everyone is highly motivated to get far more business value from their IT investment.
And that continues to be one of the original -- and most powerful -- motivations around the
whole cloud discussion.
But there are a few more motivational aspects worth sharing.
Data Is Growing Faster Than We Originally Thought
Joe shared some preview numbers from our forthcoming annual "Digital Universe" survey that we do with IDC every year.
First headline: we collectively crossed the "zettabyte barrier" in 2010, a few years ahead of schedule. We now have over a billion trillion bytes sloshing around out there. Not only is the number growing in absolute terms, but the rate of acceleration appears to be increasing as well.
Far more kinds of information are far easier to capture, store and analyze than ever before -- and there are ever-more ways to create value from it -- hence the "big data" discussion a bit later.
The forecast currently stands at 35 zettabytes within a decade. Yes, that's an insanely big number.
But, keep in mind, historically IDC has been way too conservative in their forecasts. I'm personally guessing we'll cross 100 zettabytes by 2022 -- the rate of acceleration has dramatically increased in the last few years -- the curves appear to show exponential growth vs. mere linear functions.
And it ain't your father's data types either -- 90% of it is seen to be unstructured: video, audio, pictures, signals -- and not the convenient transactional data of yesteryear -- although there's still plenty of that around when you consider these stupendous numbers.
And Lots Of Files
I particularly enjoyed the estimate that we now were collectively the proud owners of some 3 quadrillion files -- that's 300 million billion files, for those of you who prefer your big numbers in digestible chunks like I do.
A while ago, I wrote a rather controversial post "The Future Doesn't Have A File System" where I made my case that -- at some point -- file system technology can only be extended so far, and we'll have strong motivations to move to an object-oriented approach with rich metadata.
Regardless, not only do we have a lot of data, we have a lot of data objects as well to consider.
More Big - And Little - Numbers
A number I hadn't seen before spoke to the number of server images processing all this information. Simply put, IDC expects the number of server images to grow 10x this decade.
I hadn't really considered that aspect before. I guess I sort of assumed that existing server images would simply get more powerful. Sure, there'd be some server image growth, but I wouldn't have guessed 10x this decade. I'll want to dig into the details to see how they came up with that particular estimate.
That's the big numbers. The more sobering "little number" is that IDC expects IT staffing to grow less than 50% during this decade.
Joe put it well. He said "congratulations, you're all very much in demand!". But when put against the backdrop of explosive information, server and presumably application growth, those 22 million IT professionals will have to be orders-of-magnitude more productive than they are today. I suppose the corollary would be "... and you have a lot of work!"
Hence my ongoing discussion around demand for new roles, new skills and new IT org models going forward.
The New Face Of Security
I sort of resist when security statistics are used to simply scare people. Well, in this case, it might be justified. Decide for yourself. I was taken aback, maybe you will be as well.
One interesting statistic was that 88% of Fortune 500 domains had botnet activity associated with their domain, presumably coming from inside their firewall. Much like each of us carry some interesting flora and fauna on our persons, it seems that corporate IT environments have largely evolved to the same state -- most of them are already thoroughly infected with pesky critters.
I don't know if most security professionals are fully cognizant of the implications of this. One would presume that the Fortune 500 companies are collectively pretty competent when it comes to this sort of thing. But I'm starting to have my doubts ...
Going farther, 60% of this same group had their email addresses compromised with malware during 2010. So much for trusted conversations with your customers and business partners.
And it looks like the bad guys are definitely putting their hearts and souls into this new "opportunity". For example, of the 60 million known variants of malware, a stunning one-third were created in the last year alone.
If nothing else, implementing new security architectures ought to be motivation enough for re-architecting how IT is delivered. It appears we're collectively leaking like a sieve.
Taken together, there's hardly a case that can be made that existing approaches to IT security are effective.
Oh Yes, And The Business People Just Called ...
I've written before how IT looks at cloud one way, and the business side looks at cloud completely differently.
IT people want efficiency, control and choice.
Business people want speed and agility.
If they can't get that from their internal IT functions, they most certainly will go elsewhere. As many of them have started to do already.
Enter The Hybrid Cloud
A significant portion of the industry has (finally) settled down on relatively standardized terminology.
A private cloud is one built and operated solely for the benefit of an IT organization. It can reside in a traditional data center setting, or be hosted externally as needed.
A public cloud is one built that offers a portfolio of standardized services to all customers. These services can be extremely generic (e.g. commodity IaaS) or highly specialized (e.g. security analytics as a service). We use the term "service provider" to broadly describe these folks.
A hybrid cloud, then, is a rationalized and orchestrated combination of both: a mix of private cloud services and public cloud services that gives the enterprise the desired combination of economics,
functionality and trust.
Enter long discussion on how to figure out that idealized mix for your particular situation.
The Private Cloud Is Alive And Well
I tend to smirk a bit when I remember how the clouderati pooh-poohed the notion of a cloud-like environment built for the benefit of a single large enterprise. At the time, I chalked it up to them not spending enough time with larger enterprise IT organizations; hence limited empathy with the enterprise IT perspective.
Well, as we sit here today, our customers aren't talking about private cloud so much as they're actually doing it. Whether it's via an expedited path (such as a Vblock), or a hand-crafted alternative, the numbers are starting to show up in the analysts' reports, as seen here.
The ubiquitous three-phase model for internal IT infrastructure transformation is also alive and well: a signficant majority of IT organizations are now following the exact same path we laid out a few years ago.
I'm now seeing it show up in our competitors' decks, so it must be good stuff :)
The Service Provider Market Is Alive And Well
We always knew that our customers would be best served by having a large, vibrant marketplace of compatible-yet-competitive consumption options, e.g. service providers.
That trend is now well underway in the marketplace, with literally hundreds of service providers who have stood up services to target the virtualized IT customer.
Many more are coming. EMC is doing our part as well -- the recently announced EMC Service Provider Program focuses on making sure that this new class of partner is as successful as our more traditional integrators and resellers.
Hybrid Cloud: Bridging The Private and Public Worlds
Part of the challenge in considering a hybrid cloud approach is rationalizing what goes where, and why. I think we've made good progress on that front, but there's certainly more to do.
More importantly, there's a strong need to create automated control planes that span both domains. Service creation and monitoring. Information logistics back and forth. And, above all, establishing and verifying trust in this new world.
We don't have a snazzy name yet for this all-important orchestration and control function, so for now, we'll use the more generic "Automated Hybrid Cloud Infrastructure" as sort of a conceptual placeholder.
From Infrastructure To Applications
Even though the changes happening in the infrastructure world are significant and disruptive in their own right, there's much more in play.
The application world is undergoing a restructuring of core concepts that, in some regards, may end up being more impactful than the infrastructure focus of most cloud discussions.
Clearly, there's a strong motivation to move today's applications to a more attractive consumption model: whether that's done internally, externally or a rationalized combination of the two.
But what about tomorrow's applications?
Simply put, they don't bear much resemblence to the generation that preceded them. All sorts of developer-friendly frameworks are now in play: e.g. Spring, Rails, Python et. al. The focus has clearly shifted to developer productivity vs. infrastructure optimization. Rant all you want about this, the trend is undeniable.
More importantly, when done right, these newer applications are born "cloud ready" as well as largely "cloud agnostic". A surprising number of enterprises develop their own applications, or substantially integrate pre-fabricated chunks of application logic.
And their world is moving just as fast as the infrastructure folks.
Contemplating The New User?
The third part of "everything in IT is changing" can be best represented by the ubiquitous iPad. No longer are we -- as knowledge workers -- satisfied in merely having a recreation of our corporate desktops on our mobile devices.
No, we want more. Much, much more.
Indeed, I felt this topic was so important I did an extended post on the subject, which you might be interested in.
The New Security Stack
If infrastructure is changing, applications are changing, and user experiences are changing -- how should we think about topics like security and GRC going forward?
We're doing our best to simplify an inherently complex discussion, and here's our current best
effort.
First, we need to become more comfortable in thinking entirely in terms of virtual abstractions (virtual applications, virtual information sources, virtual data centers, etc.) instead of physical entities. In essence, the virtualized abstraction becomes our new "logical perimeter" for implementing controls and policies.
Policies are associated with groups of virtual entities. Those polices express desired controls for identity, information and infrastructure. As virtual entities are inherently mobile, policy (and controls) thus need to move with the virtualized abstraction.
As a tactical example: move a virtual machine, move its firewall as well as its firewall policy.
Security management then measures the effectiveness of policies and the controls: both in event
capture (think enVision) as well as advanced analytics (think NetWitness).
As we move to more externalized services, there will undoubtedly be a need for external trust authorities who can vouch for the same policies and controls (as well as event authenticity) on behalf of both producers and consumers of IT services. For example, the recent announcement of the RSA Cloud Trust Authority is a step in that direction.
Finally, enterprise-level GRC compliance dashboards will be needed to not only monitor risks in the IT world, but the non-IT world as well. The Archer platform in the EMC portfolio best represents that line of thinking.
My personal fear is that people will attempt to address the new security stack in a relatively piecemeal fashion: a little here, a little there and hope for the best. A better (yet certainly more difficult) path is to reconcieve the security architecture in light of the monumental changes occuring in the infrastructure, application and user experience domains.
Enter Big Data
By the end of EMC World, I think just about everyone understood what this whole big data thing might be about. When asked, I offer that it's a change of perspective: seeing vast mountains of data as an incredible opportunity vs. yet another IT headache.
Of course, we had the new Greenplum performance-optimized appliancess on the floor, and there was a signficant announcement about EMC offering commercialized versions of Hadoop going forward.
The Isilon team was there in force as well, showing how scale-out NAS was substantially different than the more common enterprise-class NAS we've all grown accustomed to.
Two new events were created to help us all understand just how different this world can be.
One -- the first Data Scientists Summit -- spoke directly to the end users of all this technology: who are these people, what are they trying to get done, and how can we help them?
The second event was the first-ever Big Data Storage Summit. IT at petabyte scale can be quite different, and we wanted to hear from these people first-hand what their worlds were like, and how things were changing.
It was an eye-opening experience to say the least; much more on this later.
The EMC "Big Data Stack"
Also shown at the show for the first time is our best attempt to define our views around a "big data stack" that we believe most customers will want for unlocking the potential of big data.
From the foundation, there's the need to store data at multi-petabyte scale. All storage and infrastructure operations must be designed to have as much integration and automation -- and as little human intervention -- as possible.
When the information type biases towards transactional data, that's VMAX. When it biases towards large files in a small number of locations, that's Isilon. And when we're talking millions and perhaps billions of objects that are geographically dispersed, that's Atmos.
Next level up: analysis. We've got to deal with structured, unstructured and the growing category of semi-structured data. Although we want a historical perspective, we want our answers in near real-time.
There will undoubtedly be a broad range of analytical tools that people bring to bear to leverage big data, but we think we can make a strong contribution here and now with Greenplum and now Hadoop. I'd put our deeper relationship with SAS in that category as well.
More to be done here, so please stay tuned.
Finally, insight is great, but action is better. This discussion leads you to new forms of collaboration and workflow models, where the new xCP application platform is positioned to play a key role, as well as other assets from IIG.
The EMC Extended "Cloud Meets Big Data Stack"
Joe did a nice job of putting all the various EMC technologies together on a single slide from a cloud perspective, and I think it's worth sharing.
On the left side, we've got traditional enterprise applications -- the ones we're all so familiar with. Enter classic EMC enterprise storage: VMAX, VNX et. al. That's not going to be any less important anytime soon, and there are many exciting things going on this space, not the least of which is a transition to an all-flash world.
Now, go to the right side. More and more of these applications are going to cloud-like service provider models, and we believe a new view of storage is required: enter Isilon and Atmos. In a soundbite, that's "big data storage".
All of this data has to be protected and recoverable. Look down below; that's where the BRS portfolio (DataDomain, Avamar, Networker, DPA, etc.) fit in.
In between the left side and the right side, we've got an interest set of technologies that bridge the two domains to help create hybrid clouds. VPLEX to move information. RSA to secure it. And Ionix and UIM to manage service level delivery.
At the top of the stack, we've got big data analytics via Greenplum (and now Hadoop) -- to extract meaning and insight from the data. Although there's not room on the slide, imagine xCP sitting way up at the very top, as described above.
Now, I'm not going to tell you "we're done, we have all the answers", but -- you have to admit -- it's a very promising start to a very daunting set of IT challenges.
There's more technology that needs to be added to the mix. More investment in new partnerships and alliances. More investment in creating the new skills we'll need in this world.
And, above all, more time spent with our customers and partners -- and understanding how they see the challenges ... and the opportunity.
Thank You ... From EMC
At EMC World, it's not really about EMC if you think about it. It's about our customers and partners. We exist to serve them. That's where the magic really happens.
You could almost feel the excitement at the show. Many of the people there believe we're on the verge of something really big indeed, and they're excited.
I guess I feel privileged to be a small part of all of this.
Comments