Most of AWS’s competition-companies such as Terremark, Savvis, Bluelock and their global brethren-pursue an enterprise strategy. Enterprise CSPs implement “enterprise-grade” infrastructure from Cisco, EMC, HP and their top-end competitors, and they boast about it, implying that its reliability outstrips AWS.
Generally speaking, these CSPs implement a VMware-based software infrastructure. In terms of the customer-facing offering, the vast majority eschew direct signup in favor of a high-touch sales approach. In contrast to AWS publicly available, rock-bottom pricing, enterprise cloud service providers typically pursue large-scale, multi-year deals that can run into seven figures. It’s no secret that they dismiss AWS rather sniffily, implying that a mere bookseller can’t really be viewed as a serious technology provider.
As of the end of May, it was AWS, the commodity cloud provider, vs. the rest of the CSP world, the enterprise crowd.
Without examining the merits of these two perspectives, the past 30 days have seen developments certain to shake this established dynamic and bring enormous change to it. Competition has come to the commodity end of the market, and it’s going to disrupt every participant in the market-CSPs, software vendors and users alike.
Both Google and Microsoft announced a strategy change in June. Where once they had provided only PaaS services, now they are planning to roll out IaaS services. This is obviously a big shift for Microsoft, which previously offered a .NET Framework-flavored set of interfaces. For Google, however, how big a change this shift represents cannot be overstated.
Unlike Microsoft, Google has no history of offering virtualization or operating system-based services. It previously implied that forcing developers to work at the OS level and self-manage scale and elasticity was an obsolescent approach that could now be relegated to the ash heap of computing history. For Google to shift gears and offer a virtualization-based IaaS service is enormous. Given Google’s technical certainty (arrogance, some might say), it probably was the result of a soul-searching decision-although given that it’s Google, it was probably a metrics-driven soul search.
What both initiatives indicate is that developers, no matter how theoretically superior the PaaS model that allows a higher level of abstraction, stubbornly continue to adhere to an OS-based virtual machine approach to developing and deploying applications. In a way, it’s not surprising. The most expensive capital in the world is human capital. Once skills get built, humans are loathe to relinquish them. All the developers on the planet has grown up in an OS-based world, and they’re not going to march over to PaaS overnight. They’re going to stick with what they know and love, and they’re not going to give it up for what might turn out to be a flash in the pan.
AWS understood this perfectly, launched a service aimed at this developer pool and reaped the benefit. While Microsoft and Google insisted that developers would eventually come around, AWS developed such a following that its momentum is immense-and, from the perspective of its competitors, terrifying. As one knowledgeable observer said to me: “AWS is the new Wintel, but it’s both Microsoft and Intel.” If you’re up against that, you’re right to be concerned.
1. The commodity cloud approach is legitimized.While the enterprise cloud providers could denigrate AWS as the inadequate offering of a mere bookseller, that kind of disdain is going to be much harder to convincingly put across about Microsoft and Google. With three major commodity providers&mdash’and with tons of momentum and money-the enterprise players will struggle to differentiate and gain attention. When an enterprise end user has three commodity providers to select from to host its application, how can you get on the shortlist?
2. Intense commodity cloud price competition. As the only real player in this space over the past five years, AWS has blissfully enjoyed a competition-free zone. That’s soon going to change. When multiple suppliers compete in a commodity market, prices drop quickly and inexorably. All three players have plenty of financial firepower, and they’ll bring it to bear on pricing.
If you want to know what this market will look like, consider the airline business: Cheap seats and no amenities. While many experienced flyers lament the disappearance of old-style travel with its associated luxuries-see these classic airline tv ads; the first discusses the steak AA serves!-the current market serves many times more travelers at far lower prices, albeit without any frills. (My favorite was the Ryan Air plan to charge to use the plane lavatory. Now that’s frill-free! Expect to see lots of cheap computing with a “serve yourself” approach.
3. Enterprises will struggle to locate expertise.The corollary of the last point is that adoption, implementation, migration and operation skills have to be somewhere. If the providers don’t offer them, it falls to the user. As cloud skills are in short supply, enterprises will be challenged to develop or hire sufficient talent. This shortage will be exacerbated as lower prices drive additional demand.
4. Private cloud economics will need to be re-evaluated. Many companies have justified building their own cloud infrastructure because they calculated that they could deliver computing less expensively than AWS. While those plans may have penciled out when AWS charges $.08 per hour for its smallest instance, what will the numbers look like when AWS, locked in a market share battle with Google and Microsoft, starts giving price haircuts?
Cloud Competition Means Bargain Basement Pricing What can we expect this dynamic look like in practice? We have already seen a couple examples of what happens when commodity economics start to bite.
1. Dropbox doubles storage. A couple weeks ago, I woke up and found that Dropbox doubled the amount of storage it provides for me-at no extra price. Dropbox maintained that it was responding to increased user consumption of digital media, so the new pricing supported consumer demand. A different spin came from industry observers who noted that others, including Microsoft, offered better pricing. In my view, commodity economics meant the price of online storage was dropping, and Dropbox needed to respond. Winner? Customers.
2. Redmond and Mountain View fight. In a head-to-head between commodity providers, Google prevailed in a competition to sell online office services to Dominion Enterprises. Dominion’s CIO noted that Microsoft was 50 percent more expensive than Google. Here’s the kicker, though. Google’s price was $200,000, compared to $300,000 for Microsoft. What did this new office service replace? Microsoft on-premises software that was $2 million per year. So Microsoft and Google were competing to reduce Microsoft’s current price by up to 90 per cent. Now that’s commoditizing a market.
Many people in IT draw a distinction between consumers and enterprise users. As these two example illustrate, commoditization is a powerful trend that will sweep through all markets. For those with long memories, the PC originally represented this same phenomena, and it utterly transformed both vendors and users. Cloud computing is the next platform shift-and it will have at least the same transformative impact.
Expect to see a bloodbath in the cloud service provider market as the commodity players fight for share and the enterprise players struggle to differentiate sufficiently to overcome pricing disadvantages. One thing I think it’s safe to predict: The next few years are going to be interesting times indeed.