AWS pulls in roughly $150 billion in revenue, according to Amazon financial reports, solidifying its spot at the top of the public cloud food chain. Sometimes the cloud provider misses the mark with its AWS cloud tools, though those misses don't get the same level of notoriety as its hits.
It can be particularly difficult to evaluate AWS' less popular services, as there is no real data to assess, according to Judith S. Hurwitz, president and CEO of Hurwitz & Associates, a research and consulting company in Newton, Mass. Amazon "unleashes hundreds of services every year," she said, and the model is different from traditional product-oriented software companies. New services are offered, but they only get substantial investment if they gain market traction; if they don't, Amazon moves on, Hurwitz noted.
"Not surprisingly, Amazon doesn't release information unless it's about something positive," said Carl Brooks, analyst at 451 Research in Boston. But looking at the Region Table, where all AWS cloud tools and their available regions are listed, is a good way to get some inkling of the relative success of Amazon cloud services, he added. "(Elastic Compute Cloud) EC2 is well represented across the board but some other services, less so," Brooks said.
The relatively new AWS IoT service, for example, is confined to a few regions -- currently available in two American regions and seven globally. "When something stays in just one or two regions you can probably surmise that demand didn't materialize enough to justify expansion," Brooks said. Looking at this table every six months or so could provide further insights.
"That's probably the best you are doing to get in terms of figuring out what's popular and what's not."
The Amazon business model
AWS is protected from overinvesting in a product or service by how it approaches development and commercialization, Brooks added. Each AWS tool has its own team -- as is typical among software vendors -- and the more popular the service becomes, the more the team expands. Amazon Cognito, for example, has a software team assigned to it. The team came up with the idea, developed the service and launched it.
But when a service launches, the team becomes, in effect, an AWS customer. As AWS cloud tools grow, they make use of more AWS resources -- and that boosts profitability. "They build and announce using small amounts of resources," Brooks said. "Then, if people like [the service], they grow. It is sort of failsafe approach."
Amazon Simple Email Service is one of the AWS cloud tools that Brooks believes has failed to catch on. "Notification is now probably the third- or fourth-most widely used service, but Simple Email is still in a basic availability zone after six or seven years," he said. "This is how [AWS is] protected from having a bomb. And when something grows, it grows organically. It is a self-correcting service platform," he added.
Amazon also chooses to develop offerings based on customer needs. These AWS cloud tools can originate democratically; customers ask for something like SNS, and Amazon develops it. Another way it can happen is through a request from a big-budget customer.
"We know that AWS develops against demand, and large customers can and do drive that demand, but AWS is very cagey about who those very large users are," Brooks said. AWS also develops based on general consumer demand, and AWS Elastic Beanstalk is an example of this, according to Brooks. In the last two years, AWS has catered to large enterprise customers and then "pushed the resulting service out the door to everyone else," Brooks said.
In the wings
Amazon has other development efforts that aren't immediately visible from the outside. "They've gone public with Cognito, Lambda and Inspector and various reporting tools where they are trying to broaden their appeal into existing data centers," Brooks said.
"I think Cloud9 was acquired partly because it was widely used by AWS dev teams," Brooks said. But its future could be as an element in the AWS brand, he speculated. Amazon uses another "model" to expand its footprint, Hurwitz added. "When the market is looking for a service, and they don't yet provide it, they will open up the opportunity to partners and encourage them to build something to fill that gap," she said.
This tactic brings expanded capability into the market quickly and keeps customers happy. But when AWS sees a partner service with strong potential, it will sometimes develop a competing product. "That has had some real consequences for those companies," she said.
Experts make predictions for new services
Could VR be on the AWS horizon?
AWS moves into its second decade of operation