Look for cloud computing prices to continue falling as vendors compete for enterprise IT workloads. Second in our series on cloud computing pricing.
During the first week of March, two tech giants dropped their cloud computing prices within two days of each other. First, Amazon Web Services, the market pioneer in infrastructure-as-a-service, dropped prices on its core EC2 service and slashed the hourly rate for long-term contracts. Two days later, Microsoft dropped prices on its Windows Azure compute and storage services, including a 50% cut for entry- level servers. AWS and Microsoft are up against new competitors as well as each other, a promising sign for would-be cloud buyers hoping to see prices fall further.
AWS is angling for more big customers. For example, it cut the price of a small on-demand Windows server by 4%, from $0.12 to $0.115 an hour. But it cut the price of Reserved Instances (requiring a one-year commitment) as much as 33%. Under the new deal, a Windows Server can be had for a year with an up-front payment of $2,400 and $0.264 per hour, compared with $0.92 per hour on demand. Amazon promises bigger volume discounts. If you intend to spend $5 million or more this year, "give us a call. …We look forward to speaking with you," wrote AWS chief evangelist Jeff Barr in a blog post.
Microsoft is chasing entry-level customers and developers. It dropped its entry-level, extra-small Windows cloud server from 4 cents to 2 cents an hour. That puts it on par with rivals' Linux servers, the first time Microsoft has dropped its premium for Windows.
Microsoft, with its flagship operating system and rich line of related tools and applications, is watching the Windows developer community migrate to the cloud, but often not to its Azure cloud. AWS and Rackspace have offered cheaper raw online computing power. VMware-backed Cloud Foundry offers a development platform to build apps that can deploy on a number of vendors' clouds, and VMware recently made Cloud Foundry more Windows-friendly. Hewlett-Packard, which is just entering the cloud infrastructure market, is emphasizing its own development platform.
To keep cloud app developers engaged, Microsoft must put the right resources on Azure's platform-as-a-service--developer tools, database services, and messaging services--but also make it affordable. Today's most creative new software projects often begin in a cloud, and a big reason is to keep startup costs low. Cloud computing is critical to the future of the Windows franchise.
AWS, meanwhile, needs far larger scale to make its cloud business material to Amazon's overall financial performance. It's a relatively small business today; revenue estimates range from $1 billion to $1.5 billion a year.
With their early March price cuts, AWS and Microsoft made a grab for market share and matched the price leader in entry-level cloud servers, Rackspace. Rackspace is a traditional managed hosting provider that broke into infrastructure-as-a-service by going after entry-level buyers. It was the first to offer a very small virtual server in the cloud, and for more than two years was the only purveyor of a 1.5-cent-an-hour option. AWS introduced a micro Linux server at 2 cents an hour in September 2010.
Competing On Product And Price
For the first time, these big three IaaS providers appear to be competing aggressively for entry-level buyers--which might be individual developers or small businesses or enterprise customers testing their options. Any of them could become much bigger customers.
At the same time, Microsoft has become more directly competitive with AWS by offering more powerful virtual CPUs in its Azure server lineup at equal prices. Azure's particularly competitive at the small-size server levels, offering more disk space and, in some cases, more RAM than AWS, plus more-powerful virtual CPUs.
AWS tends to offer more RAM than Microsoft does as the server instances get larger. As Microsoft focuses on small businesses, new cloud users, and departments within enterprises, it seems less committed to capturing the largest cloud customers.
AWS's push to sell Reserved Instances is a departure from one of cloud computing's main attractions: pay as you go. Perhaps AWS overbuilt and has underutilized capacity. Or its move might be a recognition that, without a lower price, the cloud can't compete with in-house data centers for long-term, predictable workloads.
Microsoft is keeping a simple, uniform on-demand pricing schedule, believing that cloud customers want simplicity, says Steven Martin, general manager of Windows Azure. He dismisses AWS's push as trying "to find more customers for underused capacity."
Rackspace's CTO, John Engates, says the company won't cut its prices in reaction to Microsoft and AWS because it set lower entry-level prices from the start.
While cloud server prices vary depending on the features, cloud storage is increasingly looking like a commodity. Google dropped its charge per gigabyte to $0.12 on March 6, retroactive to March 1. That move brought it in line with AWS, which dropped its price to $0.12 a gigabyte in February. Microsoft cut it to $0.125 per gigabyte on March 8.
Other cloud infrastructure players are emerging. Verizon bought Terremark, part of a network of VMware-blessed cloud providers, for $1.4 billion, and it didn't spend all that money to leave its new data centers idle. HP has just entered the cloud infrastructure market; it comes in emphasizing its development and support platforms, but it'll have to get to the prevailing price bar to be in the game. And then there are the cloud youngsters, including SoftLayer, Hosting.com, Layered Tech, and Carpathia Hosting.
Skepticism reigns among enterprise IT that cloud vendors can compete for big, long-term workloads. Their prices still might not be right for that work. But recent signs are that the price of compute cycles and storage in the cloud is headed in one direction: down.
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.