IaaS for Azure and inclusion of cloud use in enterprise agreements has helped Microsoft's cloud services gain momentum, says general manager Mike Neil.
(click image for larger view)
8 Things You Didn't Know You Could Do With Your iPad And The Cloud
Microsoft's Mike Neil said the Windows Azure cloud may have joined the infrastructure-as-a-service (IaaS) competition only last April, but he said the chart of Azure IaaS growth "looks more like a flagpole than a hockey stick."
Azure had been primarily a platform-as-a-service (PaaS), a developer's platform with Visual Studio and other tools available in Hyper-V and Windows Server settings. Azure IaaS came out of beta only seven months ago, while the PaaS version became available in preview form in the fall of 2009. Perhaps it's because supported IaaS has been available such a short time that its use "looks more like a flagpole" -- a spike represents both pent-up demand and a relatively small base on which to start adding customers. Now, Neil said Azure is adding 1,000 customers a day and Azure revenues are up 100% year over year; that includes the PaaS revenues as well as IaaS.
Neil is the executive who helped establish the Microsoft Hyper-V hypervisor and who is now general manager of Azure. When there was a slowdown or freeze up in Azure's staging API at the end of October, it was Neil who reported to Qi Lu, president of Online Services, and Satya Nadella, president of the Server and Tools business, on what went wrong.
Neil says that reporting process is a routine that Microsoft has incorporated into its culture. The point of it is to continuously improve Azure's uptime, not place blame on individuals. "It's not a blaming process. Our culture says an incident is a learning experience. I should fix this before it bites me in the butt. We take these learnings and translate them for customer use as well," he said in an interview earlier this month.
Mike Neil, Microsoft Azure General Manager
Implicit in the comment is the fact that Microsoft is running its own technologies to power Azure, Hyper-V, Windows Server, System Center with its Azure release pack, and its own Azure cloud software, and therefore it should be able to get the maximum out of them. It wants these components to provide highly reliable operations on the new scale of cloud computing. That scale dwarfs the enterprise operations using Microsoft that preceded it. According to documents filed in locations where it has built cloud datacenters, Microsoft invests $450 million to $500 million in a major new data center, such as the one outside Chicago, designed to hold 300,000 servers.
The more reliably it operates its cloud datacenters, the greater the chance it has to convince customers to engage in two-tiered, hybrid cloud computing through Microsoft. In fact, on the reliability front, Microsoft's record, while shorter, is as good as anyone's. One of its few major outages was the eight-hour Leap Day failure on Feb. 29, 2012. It published a full disclosure of what had gone wrong with misdated security certificates setting off a cascading shutdown of servers.
Neil said Microsoft will complete root cause analysis of its end of October service slowdowns and disruptions and publish a post mortem on it as well. "We've been pretty forthright with customers. Enterprise customers appreciate that. Our competitors have been a little more opaque," he said.
In an attempt to further engage enterprise customers, Microsoft is including whatever hours of Azure use that a customer wants in its Enterprise agreements. "That has greatly reduced friction for customers" getting started with Azure, he said. They have a predictable cloud bill stated in their overall Microsoft contract, which they prefer to a bill that dips to a $10,000 low one month and a $100,000 high the next. "It shouldn't be like your cell phone bill after you come back from Europe," says Neil, with a wry nod to roaming charges.
Multicloud Infrastructure & Application ManagementEnterprise cloud adoption has evolved to the point where hybrid public/private cloud designs and use of multiple providers is common. Who among us has mastered provisioning resources in different clouds; allocating the right resources to each application; assigning applications to the "best" cloud provider based on performance or reliability requirements.