Infrastructure 2011: 5 Trends Not to Expect

1 Summary

Just as we talk about vaporware with regard to products or projects that have yet to materialize, so too do we discuss infrastructure happenings that haven’t yet shown any real signs of taking place. I’ve highlighted several of these happenings and trends, those that get a lot of ink but haven’t been fulfilled, and won’t likely be within the next year.

Ultimately, I think that market pressure and monumental changes in IT philosophy will lead to most of these items eventually becoming reality, but it’s difficult to gauge an accurate timeline. If there’s one thing certain in IT, it’s that users are always searching for the next big thing, but that legacy vendors and equipment, as well as national technology policy, can slow down progress in immeasurable ways. With regard to corporate prerogative, I don’t know what can make Amazon and Apple open up about the details of their operations — they certainly don’t have to, and users aren’t exactly forcing openness by voting with their feet.

Infrastructure 2011: What Not to Expect:

Ubiquitous Cloud Adoption

Although there have been incredible advances in cloud computing, we’re still far from ubiquitous, or even mainstream, adoption. This coming year won’t be the year that happens, either (unless we’re talking about SaaS, of course). In all the studies, surveys (just pick one) and articles written everyday about cloud computing, the one thing that isn’t clear is that corporate users understand there’s a process to cloud adoption. Often times, it’s from the bottom up (i.e., it starts with developers and ultimately earns CIO blessing) or involves testing the waters by hosting relatively low-risk and unimportant applications in the cloud. No one’s actually buying infrastructure as a service, everybody is planning to do so, but what about test-dev or ad hoc projects like special web sites? When the majority of businesses start even experimenting with cloud computing or using it for low-risk projects, then I’ll believe widespread, meaningful usage is on the way.

Amazon Will Announce its AWS Revenues

Let’s be honest, all anybody wants to know is how much money Amazon makes from its cloud services, and Amazon isn’t talking. 2011 won’t be the year we find that out, either. Perhaps it’s waiting until the business hits the billion-dollar mark, or perhaps it just doesn’t want to give the competition any inkling as to how they compare, but mum’s been the word thus far, and it doesn’t appear AWS is any worse off because of it. Maybe Amazon just likes watching reporters and analysts speculate as to how much money it’s making from cloud computing. Even the most-thoroughly researched estimate thus far, done by UBS, showed profit margins beyond what most pundits thought were possible. If I’m wrong and Amazon does finally disclose its cloud revenues this year, they’ll blow industry watchers away.

Apple Will Tell Us Anything About Its Data Center Operations

It’s remarkable that in a technology world enamored with open source and general openness, Apple, with its very closed approach, is the biggest thing going. Now that Apple is running a massive data center and working with Hadoop, I expect that approach to carry on into operational openness — we shouldn’t expect to get any details as to how it’s managing airflows or boosting performance across thousands of nodes. Even now, after all the reporting about its billion-dollar iDatacenter in North Carolina, nobody really knows what it’s being used for. Apple is wading into the massive web-player world of Google, Yahoo and Facebook, but there’s no way it shares the operational and infrastructural secrets to its success the way those companies have. Maybe that’s because Apple is plotting a big move into the cloud services business, and any information about what it’s developing to solve its problems will tell altogether too much about what new products are on the horizon. Unlike web companies that rely on developer communities to improve their offerings and, thus, release early previews and offer up some code, one secret to Apple’s success appears to be its tight-lipped approach to new products. Whatever it’s working on, that first-mover advantage will play a large role in helping Apple drive hype and define (or redefine) the market.

Legitimate Progress on Clean IT

It’s easy to pick on Facebook for not choosing data center locations where it can rely on renewable energy, but it’s not Facebooks’s fault. It’s the fault of governments, if it’s anybody’s fault, and it seems unlikely that too many state or federal legislators will, in this political climate, push the issues that need to be addressed to solve our energy situation. What’s needed are economic incentives for those building data centers to implement solar or wind generation, or for an energy infrastructure that can transmit clean power from where it’s produced to where it needs to be used. Until those with the power to do so decide it’s time to spend the money to make clean energy happen, advanced cooling techniques and low-power servers are the best we can ask for. They’re laudable efforts, just not ideal efforts.

The Demise of Intel

No matter how popular ARM or other new server architectures become, Intel isn’t going anywhere. And it’s not just because Intel has such an overwhelming presence in servers already, as simple math dictates that if customers start buying new processors en masse, they’ll certainly be buying fewer Intel processors. The reason is that Intel almost certainly has a plan in place to address this demand if it actually comes to fruition, even if it hasn’t disclosed it. As long as there’s money to be made from selling x86, Intel will push it first and foremost (it has that market cornered, after all), but there’s no way it hasn’t done work to develop alternative architectures, or hasn’t thought about acquisitions (e.g.,Tilera) that would let Intel buy its way into that space. If the name of the game is low power, not entirely new architectures, Intel likely has a server-ready Atom processor ready to go, which it could push to traditional server vendors and expand the Atom-based server market beyond SeaMicro. It’s natural to get excited about newcomers with grand visions, but — in the near term, at least — it will take an act of God to displace Intel from its processor throne.