"Each of these trends is powerful on its own and requires devoted study from IT planners and leaders. Taken collectively, they can be a rather large tsunami that engulfs organizations faster than they can adjust."

Opening Statement

What a difference a few years makes! At the end of 2008, we were entering the depths of the great recession, and everyone in IT had battened down the hatches, shuttered the windows and doors, and scrambled for their shelters to wait out the storm. Innovation was erased from the CIO agenda, and just as prior downturns had done, this one prompted a collection of pundits to question the viability of the role of the CIO and the modern IT organization. Rome was burning. The end was near.

But was it? Appearing unaware of the surface features of a changing economy, innovations in IT pressed onward just as they had done in the decade prior. All things hardware continued to get faster, better, and cheaper, with some items moving up this curve more quickly than others. Now we are seeing some of these trends bearing fruit. Here at the beginning of 2012, things are decidedly different than just a few years ago.

Why is this so? Aside from the slow climb out of the great recession, the continued evolution of computing's three fundamental components is beginning to reveal the next set of IT disruptions. These components are central processing units, storage (transient and persistent), and networking (communications), and four of their main effects are the rise of social media, the consumerization of IT, the emergence of "Big Data" analysis, and the oncoming crush of cloud computing. If you step back and view the larger mass of computing, the incessant and differential evolution of the three components is conjuring up the ghost of the mainframe past and is about to give it flesh and bone -- but now in a wholly different skin. To see why this is, let's review recent developments in the three components of computing.

WHAT'S HAPPENING WITH THE THREE FUNDAMENTAL COMPONENTS OF COMPUTING?

1. Central processing units. CPUs are continuing to get faster and denser, largely through the manufacturing of CPUs with more cores. Some are arguing that silicon may be finally at its limits, and research into new approaches (non-silicon-based computing, quantum computing, etc.) provides glimpses of what might be. Either way, for the next year or so, silicon will continue to increase the speed of basic information processing.

2. Storage. Within the ranks of storage, we are seeing the first break from computing's dominant paradigm. Since the dawn of computing, we have been at the mercy of slow spinning disks and have established elaborate architectures to deal with the disparity of performance between CPUs and memory and these slow disks, which have failed to keep pace with the price/performance advances of processing and memory. As memory continues to "shrink" in cost relative to disks and the attendant networking gear in front of them, more and more vendors are dropping spinning disks from their architectures outright, or they are adding faster and faster CPUs that sit on top of spinning disks and direct their activities more intelligently, with compression and automatic movement to and from higher- and lower-cost storage architectures. Either way, disks are starting to take a back seat to faster and cheaper memory in terms of driving innovation for applications.

3. Networking. Finally, networking continues to make advances, not so much in speed as in the rapid commoditization of 10 gigabit circuits and what will likely be an even faster commoditization of 40 gigabit circuits. All of this will improve the price/performance of networking further. This means that we can connect collections of CPUs with transient and persistent memory stores across larger distances and with greater throughput. To use an automotive metaphor, with memory, CPUs, and fast connects between nodes of CPUs reaching new price/performance levels faster than spinning disks, vendors and architects are finding ways to get more fuel (data) through the engine (CPU) via better fuel injection and bigger exhaust pipes (high-speed interconnects between nodes and 40 gigabit circuits between clusters of nodes).

In a sense, we can now build very large systems, spread across a large geography, utilizing many kinds of end devices with better integration across most of the essential parts (CPUs, memory/disks, and networks). It is as if the mainframe is trying to bring itself to life again, but this time encompassing far more users, devices, and scale than ever before. After years of being maligned, the mainframe is back, and it is very, very angry.

IN THIS ISSUE

Advances in the components of computing have enabled massive social media sites and application tools like Facebook and Google. It is the scale of these platforms that makes social networking interesting and economically viable, politically potent, and of great interest to businesses. As Cutter Fellow Steve Andriole and Cutter Senior Consultant Vincent Schiavone point out in their leadoff piece on social media, this new form of computing is powerful because of its massive reach and its massive volume of data and interactions. With social media now so globally pervasive, the authors pose a simple question: why wouldn't a company invest in social media, especially in the analytics of social media? They deftly delve into the critical elements of a social media strategy, how companies ought to take advantage of this technology, and how the future of social media may unfold. Here in 2012, any company that thinks social media is a passing fancy or isn't relevant to them simply hasn't been paying attention.

As CPUs and memory have increased their price/performance profile more dramatically than storage and networking, and with the advent of 64-bit operating systems that can address multiple cores of memory and multiple nodes, it was just a matter of time before business applications exploited these new capabilities. Many software vendors and the open source community now produce analysis tools that can process large amounts of data very quickly. The industry is affectionately calling these "Big Data" tools.

Our second author, Thornton May, pokes holes in the term "Big Data," pointing out that it has no clear definition and that the focus on "big" may miss the mark. May makes clear that the variety of types of data, the variation in analysis methods, and the speed of business decision making should also be considered when looking at the Big Data trend. He then goes on to give several current, real-world examples of organizations that have successfully used Big Data methods for business advantage (and one that didn't).

Where this goes over the next few years ought to be interesting. Supercomputing and high-performance computing (HPC) architectures are now being regularly used for Big Data problems, and cloud providers are offering HPC architectures for short-term use. Big organizations that can get their hands on large amounts of data, especially very large IT consumer technology companies (telecommunications, social media companies), financial services companies, and governments, are likely to find the organizational wherewithal not only to collect massive amounts of data on all of us, but also to use this data -- for good or ill.

As Andriole and Schiavone pointed out with regard to social media, the consumer technology market now involves billions across the globe, and its leading device, the smartphone, has taken center stage. Here we can see how all three fundamental components of computing (processors, memory, and network) and their progression enable this trend. Back in the 1980s, BYTE columnist Jerry Pournelle predicted that in the future there would be at least one CPU per user. At the time I read that, it sounded fantastical, but Pournelle's prediction was absolutely right.

Likewise, our third author, Cutter Senior Consultant Jim Love, is correct in pointing out that the tension between the current consumer market and the resistant IT organization will be a challenge in 2012. Love argues forcefully that we in IT should not resist this change but embrace it. If we think more broadly about what is going on, it becomes clear that cell phone data networks are continuing to grow faster (although sometimes not fast enough); are being augmented with WiFi networks at home, work, and play; and are being used to reach hundreds of millions of user devices with powerful CPUs distributed across the globe. With the ubiquity of Internet protocols and with HTML5 poised to be a common-enough user interface development environment, consumer devices and software can now reach unprecedented scale at lower costs than ever. Love writes that the "tipping point" for consumerization is near. I would argue it is past us already, and the rest is merely denouement. Laggard IT shops will have to catch up to those organizations that sprinted ahead.

In our final article, I take on the entirely obvious trend of cloud computing and discuss the different ways it is likely to impact organizations in 2012 and beyond. Once again, if we think about what has been happening to computing's three components, it becomes clearer that cloud computing is a logical consequence. Why? Commodity-class, large-scale clusters are already commonplace, Web services have been maturing over the past decade to allow for better integration, and server virtualization has matured, with three or four dominant ad hoc standards emerging (hypervisors from VMware, Microsoft, Citrix, and KVM). Large-scale and reasonably well-integrated infrastructure has now been made available as a commodity product on demand. As regional networks achieve 40 gigabit (and beyond) speeds, cloud providers and cloud consumers can build virtual data centers spanning multiple data centers across the globe.

The commoditization and good-enough standardization of these technologies can now begin to have some significant impacts on what IT leaders think and how IT organizations perform their work. Rather than focus on building data centers and storage racks months and years in advance of demand, they can shift to just-in-time procurement of infrastructure, potentially increasing agility and cost-effectiveness. The pricing and procurement models that exist today in the cloud market and new models likely to emerge (spot markets, futures contracts, and perhaps option contracts) will enable a new level of risk management and prediction that can radically change the IT leader's role, making it more about managing the total cost of information and less about sweating the details of IT infrastructure.

TO TAKE ARMS AGAINST A SEA OF TRENDS ...

Each of these trends is powerful on its own and requires devoted study from IT planners and leaders. Taken collectively, they can be a rather large tsunami that engulfs organizations faster than they can adjust. 2012 will likely emerge as a critical transition year in which two well-established trends, social media and consumer technology, meet up with two emerging trends, Big Data and cloud computing, and transform corporate IT from the inside and the outside. There is a lot of promise and peril in these trends. To some, they represent the beginning of the end of the modern IT organization. I -- and I think our authors -- would disagree. This is merely the end of the very early phase of an important transformation that will require all the best and brightest IT leaders across the globe. New tools enable new strategies, and, boy, do we have new tools available to us in 2012! To sort this out, I encourage you to read on and learn how these trends will impact you and your organization.

ABOUT THE AUTHOR