My last post noted that the IT industry appears to suffer from cloud computing ennui, as the number of Google searches for the term over the past two years has dropped significantly. I also said that other evidence indicates that many IT users appear to have put cloud computing in the “done and dusted” category despite not really understanding it very well.
I went on to say that, far from being completed, cloud computing growth is accelerating. The biggest challenge to IT is that end users, frustrated with the leisurely pace of internal IT group cloud implementations, will solve their problems by placing applications in public cloud computing environments.
Predictably, the comments disagreed with me. One took a blase attitude toward the embrace of public clouds, citing the repeated “outsource, than insource” waves within IT and suggesting we’ll soon see a move away from public cloud computing. Another viewed public cloud adoption as a short-term dalliance, as “it’s been proven that in many cases public clouds are more expensive TCO than an in house or a rack in a colo, especially for relatively steady loads.”
In other words, public cloud computing is a trend destined to die out, or at least be limited to atypical use cases, with the preponderance of application deployments following the time-honored practice of internal data center placement.
Last week, at the Cloud World Forum in London, I saw how this public/internal struggle plays out in the real world. It was an eye-opening experience. One thing I like about CWF is its heavy emphasis on case studies presented by end users. I attended three presentations that show how companies embrace public cloud computing to solve problems that traditional IT practices can’t address.
Lotus F1 COO Patrick Louis discussed his team’s embrace of cloud computing to support its race calendar. While I was aware of the extensive telemetry data race cars produce, I hadn’t appreciated how dynamic an F1 team’s cars are during the year’s schedule of races. Not only must the car be modified for each race course’s unique configuration, the modifications are performed during practice runs and even during the race, based on the conditions of the specific race (temperature, humidity, and so on).
Lotus requires a huge amount of historical data as well as real-time data to respond to these conditions and makes its car as competitive as possible. In the past, this meant Lotus transported large amounts of equipment to each site. Now it’s using public cloud computing to increase capacity and reduce costs.
Lonely Planet Online Platform Manager Darragh Kennedy described how the travel guide publisher shifted its Web presence from an internal data center to a public cloud environment to better manage erratic traffic patterns and support more frequent application updates. Lonely Planet now runs its entire infrastructure in a public cloud environment.
Lest you think only avant-garde technology companies or bohemian travel firms embrace public cloud computing, John Pillar, head of software engineering for Marks & Spencer, presented how M&S used integrated multimedia (TV adverts, Twitter, Facebook and Web-based voting) as its key Christmas marketing vehicle. The campaign asked people to choose the name of celebrity endorser Rosie Huntington-Whiteley’s terrier puppy, kicking off with a lengthy fairy tale-inspired TV spot featuring Huntington-Whiteley and Oscar winner Helena Bonham Carter that directed viewers to a campaign-specific website to vote for either “Magic” or “Sparkle.”
M&S placed the website in a public cloud environment to ensure sufficient capacity to handle application traffic. In the end, more than 130,000 people voted, with the result announced during a broadcast of the very popular Downton Abbey. (Sparkle was the people’s choice; here social media experts discuss the Marks & Spencer campaign.) M&S, by the way, is no arriviste; it’s been around since 1904.
These examples show that organizations large and small, young and old, embrace cloud computing – despite the conventional wisdom that the cloud is really nothing new, or will eventually die off due to its unsupportable economics, or IT’s insistence that cloud will make sense once it builds out an internally hosted cloud environment.
Competitive Conditions Send Companies to Cloud
Given this perspective, there’s an obvious question: Why did these organizations turns to public cloud environments, despite their putative disadvantages? I believe the answer is clear. Competitive conditions require faster delivery and innovative solutions that existing IT environments and processes couldn’t support.
F1 racing is one of the most brutally competitive business environments around. Louis notes it’s a global industry worth $4 billion. Lotus measures its competition in meters per lap. In such an environment, being able to rapidly react by gathering data and running simulations is critical. Cloud’s elasticity and enormous capacity accelerates this function, letting Lotus stretch for the minute improvements that make the difference between spraying champagne on the winner’s podium and disconsolately packing up the race trailer as an also-ran.
Lonely Planet, meanwhile, faces a changing set of competitors. I asked Kennedy why his company had instituted its changes. His competitors used to be other book publishers, with product cycles measured in months or years; today he competes with Google and online travel sites that roll out constant updates. The ability to update information or roll out new features almost immediately is vital.
Finally, it’s no secret that Marks and Spencer operates in a difficult industry. It faces brutal competition and an ever-changing list of competitors. M&S also occupies the perhaps most difficult segment of the industry: The midmarket. Today’s customers can find low prices at downmarket retail outlets and online ecommerce companies, while upmarket buyers have high-end retail shops to serve them, so midmarket stores fight over a shrinking set of buyers. Finding new ways to engage and service customers is critical to M&S’s future, and its traditional IT environment proved inadequate to the task.
In all three cases, pressing business conditions forced these organizations to turn to public cloud environments, even though each already had existing IT environments. Each said its existing environment couldn’t support the IT capabilities it needed to compete in the marketplace. What the firms needed to do could only be delivered by public cloud computing – Amazon Web Services and Microsoft Azure in these examples, perhaps illustrating why Gartner’s most recent IaaS Magic Quadrant shows only these two in the coveted Leader quadrant.
Cloud Adoption Also Means Agile Adoption
Interestingly, all three organizations changed their processes as part of their public cloud adoption effort. Each moved from a waterfall application lifecycle to an Agile/DevOps mixture, as illustrated in this pair programming effort at Lonely Planet. (Trio programming, actually.)
Lotus implemented DevOps in operating its race systems to facilitate rapid simulation execution. Marks & Spencer appears to have moved to a “NoOps” approach; application developers also deploy and support their apps.
One can see, then, that embracing cloud computing isn’t merely a matter of using a different deployment location for applications. It requires rethinking established processes designed for a different environment so that infrastructure agility is matched by a streamlined deployment process. None of the companies I’ve discussed could have succeeded by trying to retrofit manual processes designed for yesterday’s infrastructure onto an agile cloud environment; process friction would have prevented them achieving the speed and innovation their business conditions required.
In a way, one could consider this need for intertwined infrastructure and process change an existential threat to traditional IT, since it necessitates a fundamental rethinking of established operating assumptions. The attitudes reflected in the comments highlighted at the top of this post indicate that many people hold a strident belief that business as usual can address the factors motivating cloud adopters – but these examples show how wrong they are.
Perhaps the most striking – and disturbing – aspect of Cloud World Forum was the disconnect between what the case studies and the vendors on the show floor presented. From what I saw among exhibitors, it could have been 2009, with cloud computing still considered an emerging technology requiring a cautious approach. The gallery below contains images showing examples of themes presented by many vendors: The incrementalist “enterprise” approach so beloved of cloud skeptics.
Given the messages put forward on the show floor, it’s understandable that many IT practitioners believe that we’re still in the early days of cloud computing, with plenty of time left to formulate a plan to address those pesky application groups. The only problem with this response is that, despite the show floor theme, we’re way past the “figuring it out” stage. As the case studies describe, real companies are solving real problems with cloud computing – today.
IT doesn’t have the luxury of a leisurely response to public cloud computing. It’s put up or shut up time if you want to continue to be responsible for running your company’s technology environment.
Named by Wired.com as one of the 10 most influential people in cloud computing, Bernard Golden serves as vice president of strategy for ActiveState Software, an independent provider of CloudFoundry. He is the author of four books on virtualization and cloud computing, his most recent book being Amazon Web Services for Dummies. Learn more about him at www.bernardgolden.com.