Unlike the early, uncontested days of public cloud, AWS faces a wide range of competition as it tries to build its IoT business.
AWS IoT, a platform to store, analyze and process data from internet of things (IoT) devices, includes the underlying heft of the most dominant cloud infrastructure available and a purpose-built architecture for delivering the nascent service. But despite some early praises, Amazon must continue to innovate and potentially tailor the service to specific industries, as a host of other massive corporations vie for this emerging market.
AWS faces a big and diverse set of competitors in IoT, including familiar faces such as IBM, Microsoft, Oracle and SAP, but also industrial and manufacturing heavyweights such as GE and Siemens. "In this space, it's not just the traditional competition of Amazon versus Microsoft; it's Amazon against Microsoft, GE, Bosch and so on," said Paul Miller, senior analyst for CIOs at Forrester Research.
Amazon also must deal with coopetition -- exemplified by Salesforce's decision last month to build its IoT platform on AWS, built from scratch but without using Amazon's purpose-built IoT platform.
IoT is a different challenge than what Amazon has faced before.
"In the public cloud space, they had multiyear advantages," Miller said. "On IoT, they're starting more or less at the same place as everyone else."
One of the early beta users of the platform was Rachio, a Denver-based seller of smart sprinkler controllers. Rachio used Electric Imp to connect its first-generation system to the cloud, but wanted to control the "last mile" of connectivity as it developed the next version.
Rachio built a prototype plug-in with an MQTT broker linked to Amazon Simple Queue Service and Simple Notification Service, but this presented huge problems with security, scaling, monitoring and high availability, said Franz Garsombke, Rachio co-founder and CTO. Amazon approached Rachio about being a beta tester for AWS IoT, which solved those major pain points the plug-in created.
"At least 40% of our development would have been spent on those things and there would be zero value to the business working on them," Garsombke said.
Franz Garsombkeco-founder and CTO, Rachio
Rachio rolled its own platform using HiveMQ until AWS IoT was ready for beta. But AWS provided the most seamless connection, Garsombke said. Ninety-five percent of its infrastructure was already hosted on AWS and its developers are trained on the public cloud -- something that could be a leg up as Amazon tries to retain its own customers as they start to look at IoT.
"That's what makes [AWS] attractive -- you don't have to cobble together different vendors or services," Garsombke said.
Improvements have been made since the product was available to a more limited audience, he added. In beta, everything was done through a command line for creating rules, editing and managing, but the console that's now incorporated with the platform makes it easier to manage everything holistically.
IoT verticals and a work in progress
Amazon has a fairly broad set of services to address the market, including DynamoDB, Lambda and Kinesis. It also has a global reach, a growing partner ecosystem for onboarding and integrating edge devices, and a vision to evolve the architecture and services to keep up with customer demand, said Alfonso Velosa, IoT research vice president at Gartner. But questions remain about the integration of the underlying services, the overall user experience, and the breadth of the software development kit and training certificates.
"It's a work in progress," Velosa said. "Those are things you can expect [to improve], but it's not completely there today."
The biggest weakness of the platform is the lack of vertical integrations, Velosa said. GE, for example, leans heavily on its industrial experience to sell its Predix cloud directly to companies in that same market, while IBM Watson IoT includes flavors for vehicles, buildings and retail.
Most enterprises are not looking for a vanilla service, whether it's to control a building or facilitate the needs of a healthcare provider, he added. There always will be alpha enterprises that want to do it themselves, but most IT departments lack the skills or time.
"They need something preferably shorter to market or something that gets them to look like heroes as soon as possible," Velosa said.
Improving the platform is more about incremental steps than glaring holes, including opening up some more APIs, Forrester's Miller said. Microsoft's Azure IoT Suite is closely tied to the company's analytics tools, but Amazon likely will do more to integrate analytics into its IoT platform in the months ahead, he added.
Predictive analytics could provide alerts when something is off, such as a pattern of irregular messages from a specific controller, Garsombke said.
"I'd like to see something where they can start pushing the outliers or information that would be really hard for me to find out without culling through the data," Garsombke said.
Despite some early shortcomings, few doubt Amazon can succeed, given its track record with AWS. It's even melding its consumer side with its cloud business in ways previously unseen. Amazon Echo can be used along with AWS to build IoT applications and, last month, a new programmable version of the Dash button sold out in just a few hours, though it's unclear how many were made available.
Since most enterprises are still in proof-of-concept stage or are just starting to move to pilot programs, there's still time to ramp up capabilities while being mindful of the competition.
"The mega vendors, including Amazon, will not be truly relevant until 2018," Velosa said. "It will take a bit more time to coalesce their packaged solutions to really hone in on the customers."
Trevor Jones is a news writer with TechTarget's data center and virtualization media group. Contact him at [email protected]
Where Azure stands in the IoT battle
Making the connection in IoT on AWS
Take baby steps into IoT and big data