This has been a curious year for cloud computing. The technology has moved into mainstream consciousness, but many vendors remain frustrated with the pace of enterprise adoption. While widespread agreement about the importance of cloud computing is present, many vendors see enterprises pursuing internal cloud implementation projects with a slow pace. As you can imagine, vendors are impatient with this pace—but not as frustrated as early-stage investors in those vendors.
Notwithstanding, I expect 2013 to be an inflection point for cloud computing, although not in the way many IT organizations or vendors do. You can expect that cloud computing trends of 2012 will become more vivid in 2013 and will prove disconcerting to incumbents, no matter which side of the vendor/buyer table they sit on. Cloud computing will prove more disruptive to the established order of things than almost anyone anticipates, and it will prove to be extremely uncomfortable for many.
Here are five things to look for in 2013.
1. The Phrase 'Enterprise Cloud' Goes From Market Prediction to Drinking Game
This year, I've heard vendors make countless number of confident predictions about the importance of the "enterprise cloud"—once enterprises realize that their old trusted vendors have bright shiny cloud products in stock, they'll start buying like mad.
The sotto voce in this is the belief, or hope or wish, anyway, that presenting customers with an "enterprise" offering will then curtail their dalliance with Amazon Web Services. It's like a spurned spouse waiting for the lure of home cooking to bring their wandering partner back home.
This will be a put-up-or-shut-up year for the attractive powers of enterprise cloud computing. December 2013 will find many incumbent vendors standing in the doorway, listening anxiously for the footsteps of wayward customers. By the end of the year, people will recognize that just repeating the phrase "enterprise cloud" is insufficient as a vendor strategy.
It's critical to understand that differentiation exists in the mind of the buyer, not the mind of the seller. Cloud service providers will need to understand the motivations of buyers and will also need to understand that many cloud buyers are different from traditional enterprise buyers—they may work in enterprises, but they don't have the same needs or goals as their counterparts. Proffering a product with the traditional functionality, language and selling approach to this new type of buyer is pointless.
In 2013, I expect "enterprise cloud" to be a meme that moves from a market segment description to a drinking-game meme. Every time you hear it, you have to take a sip of your beer.
2. Enterprises Realize That Cloud Computing Means, Well, Cloud Computing
One big realization on the part of enterprise IT groups this year has been that business unit adoption of public cloud computing isn't going to go away. Many, however, continue to implement internal clouds that are little more than warmed-over virtualization, with the expectation that once the internal offering is available, it will become the default choice of developers.
Next year will bring the recognition that developers have embraced public cloud computing because of its speed and agility. Enterprise IT groups will then comprehend that internal offerings have to meet the much higher expectations of developers—expectations raised by what they now recognize is possible. Internal IT groups will confront the reality that warmed-over virtualization isn't nearly enough—and getting to public levels of agility will require process streamlining and end-to-end automation. In a word, it will mean re-engineering, which is much more difficult than buying a new product with the expectation that it will magically make one's private cloud as attractive as the public alternatives.
Expect to see many, many articles next year on the organizational change necessary to successfully implement private cloud computing. If you're implementing a private cloud and haven't thought through how you're going to integrate end-to-end automation and remove process roadblock, you have a problem. A big problem.
3. The Public CSP Business Will See a Pricing Bloodbath
The recent tit-for-tat storage price battle between Google and Amazon, key to the first AWS Re: Invent conference, is just a warm up for 2013. Next year will see ferocious price competition as CSPs attempt to blunt Amazon's growth. Even those that have heretofore eschewed price competition will have no choice but to jump into the fray.
For example, Hewlett-Packard, which not that long ago said "the notion of just standing up a VM for raw compute is kind of done" and promoted the importance of a cloud ecosystem, this week launched its OpenStack-based cloud service with a pricing structure that undercuts Rackspace by a third. That's not exactly the mark of a provider focused on value-added services.
We'll see plenty more of these price wars as CSPs recognize platform wars are a market share land grab. The winners turn into monopolies and the losers slink off the field. However, as Warren Buffett so memorably says, "Only when the tide goes out do you discover who's been swimming naked." Translation: pricing wars are going to show who's really prepared to be a volume player in cloud computing.
Next year, cloud providers will come to understand that cloud pricing is a marginal cost-yield management exercise; efficient design, low-cost operation and, crucially, high utilization are fundamental to success. Land grab economics favor CSPs with access to significant capital. Expect those with access to pursue that advantage. It's no accident that Amazon recently sold $3 billion in bonds. It's arming for the battle.
One can expect real dislocations in the CSP industry as players have to rethink their business plans based on lower revenue streams, including some who conclude that being a cloud provider is a never-ending money pit and decide to exit the industry. We've already seen a couple of departures: the high-profile GoDaddy and the lesser-known Harris, which exited after reportedly spending $100 million on a cloud infrastructure.
They won't be the last. Look for a real bloodbath in the CSP industry next year—but realize that, for customers, this bloodbath will pay enormous dividends in lower costs.
4. Hybrid, Heterogeneous Cloud Computing Comes to the Fore
Next year will bring home the fact that enterprises are never going to settle on a single cloud technology or provider. By definition, enterprises are complex, heterogeneous technology environments. They collect multiple products in a given technology category the way others collect bottle caps or old magazines: Randomly and irrationally. Even if a given enterprise resists the urge to buy one of everything, its nice, neat, homogenous environment gets spoiled the instant the CEO decides to buy another company that inevitably, standardized on a different product.
Many IT organizations and CSPs have hoped that the world would standardize on a single technology—most commonly, one based on VMware—but that's just not going to happen. The reality for every enterprise of any size is that it will use multiple cloud technologies spread among multiple deployment environments, with those technologies all providing different orchestration software frameworks, heterogeneous APIs and single-purpose management infrastructures.
The task for enterprise IT for 2013 is to comprehend this fact and develop plans to implement a management framework that can span all cloud environments in use. Key requirements include consistent identity management, common monitoring, a management platform that provides a single pane of glass to control all cloud environments and common billing (more on that in a moment). Expect to hear lots in 2013 about cloud management or cloud broker products, which sit above multiple cloud environments and provide these key requirements.
Just as Esperanto has the beauty of logic but fails in the reality of a polyglot world, so, too, will the fantasy of a single cloud technology environment used everywhere confront the messiness of the typical enterprise IT world. Once one accepts life will never live up to an idealized vision, then one can get on with the real job of dealing with things as they really are.
5. Cloud Spend Management Comes Down to Dollars and Sense
Next year will also bring home the financial implications of enterprise adoption of pay-as-you-go pricing. The negligent habits typical of enterprise system management, in which utilization rates still commonly hover in the mid-teens, carry heavy financial burdens in today's cloud world. The upfront capital investment practices of traditional IT paper over scandalous financial practices—can anyone imagine running an automobile factory at 15 percent utilization?—but the sunk cost and lack of ongoing payment pain make it easy to ignore waste.
Not so in public cloud environments, where every month brings a reminder of the cost of running systems. Many companies are now spending tens of thousands, or even hundreds of thousands, of dollars per month with AWS. Those numbers draw the attention of upper management. Questions will inevitably arise: "What are we doing to spend so much? Is there some way to detail what we're spending and whether we're getting our moneys worth?"
While Amazon does its best regarding billing, enterprises running complex application environments—dozens of applications, deployments for development, testing, load testing, staging and production for each app, and multiple versions of each application—clearly find the current state of AWS billing inadequate.
Look for 2013 to be the year of cloud spend management solutions to become a big deal. (To be clear, these are systems designed to manage the amount of money you're spending on running your cloud-based applications, not cloud-based spend management applications such as Ariba.) These systems track what you spend, identify under-utilized servers and make recommendations on different deployment arrangements that can save you money. They don't help you operate your applications to be elastic in the face of variable demand—that's the job of cloud management software—but they can help you make sure you aren't wasting money on overprovisioning servers or not choosing spending plans wisely.
Overall, 2013 Will Be a Year of Transition, Change
We are in the midst of a multi-year shift from traditional static compute environments based on inefficient asset ownership to a new form of agile environments based on infrastructure rental accompanied by careful operation. Next year will be the year in which this transition reaches a point where the discussion shifts from "It's not like what I've currently got and doesn't have the advantages my current environment brings" to "It's not like what I've got, and thank goodness, because it doesn't come with all the drawbacks my current environment has."
In every innovation, at a certain point the weight of the evidence becomes overwhelming. Early adopters are vindicated, mainstream users are convinced and diehards fall back into the distance, their cries of "It's not good enough!" dwindling to an inaudible whimper. Every new platform solves the shortcomings of the previous generation platform. Every new platform also brings its own drawbacks.
Cloud computing is no different. It's powerful and innovative, but it has blemishes of its own. However, 2013 will be the year that we start addressing the real drawbacks of cloud computing—the challenges of scale, complexity and change—rather than fixating on its supposed drawbacks such as security, compliance and SLAs.
Bernard Golden is the vice president of Enterprise Solutions for enStratus Networks, a cloud management software company. He is the author of three books on virtualization and cloud computing, including Virtualization for Dummies. Follow Bernard Golden on Twitter @bernardgolden.