Complex pricing is making it impossible for enterprise IT departments to manually compare the cost of using Amazon, Microsoft and Google’s cloud-based internet of things (IoT) platforms, claims 451 Research.
Owen Rogers, director of the market watcher’s Digital Economics Unit, said the lack of consistency and clarity in how the big three cloud firms price their offerings is making it difficult for enterprises to carry out cost comparisons.
“I would go as far as saying it would be impossible for an enterprise to really understand the pricing implications for IoT as it is now,” Rogers told Computer Weekly.
“Each of the platforms has a completely different pricing model and their own nuances. There are so many variables, and often these variables are tucked away.”
An example of this is how they qualify the type of data that passes between the IoT devices the platform, which – in turn – dictates how much users should be charged.
Some devices send small amounts of data to test whether or not their connection to the IoT platform is still live, and some providers charge users for those whereas others do not.
“Enterprises will need to understand all of the pricing and all of these hidden terms and conditions, and even then – if they understand all of that – they’d need a perfect understanding of how much [data] they are going to consume in the future, and also their use case,” he said.
To emphasise just how complicated a job it is, Rogers said 451 Research has stopped trying to manually track the market itself, and now relies on machine learning-based Python script to carry out its cost analysis research.
“We came to the conclusion that there are nine things that ultimately affect how much each of these [cloud IoT] platforms will cost, and this piece of code will simulate a random pricing scenario based on these nine parameters,” he said.
In total, 10 million different scenarios were modelled using this approach, leading 451 Research to conclude the Microsoft Azure IoT Hub platform is the cheapest to use at scale, while Amazon Web Services’ (AWS) IoT Core offering represents better value for money for smaller deployments.
“Most enterprise scenarios today are within the AWS sweet spot,” said Rogers.
The Google Cloud IoT Core offering, however, was never found to be exclusively cheaper in any of the 10 million modelled scenarios, but did come in second cheapest in some instances. This is despite it having the simplest pricing model overall.
While the cloud giants have a history of standardising their individual pricing models to bring them into line with one another, Rogers said this is unlikely to happen in the immediate future.
“We essentially did the IoT study now because a lot of the pricing recently changed, suggesting it won’t change again very soon. Azure, for example, cut its IoT pricing by 50% last month, AWS recently came out with a new pricing model and Google only launched its service last year,” he said.
“Pricing is more likely to become standardised on services that are beginning to become commoditised. Look at virtual machines – they’re fairly similar between cloud providers, whereas IoT is still a very experimental [use case], so the need to be price clear isn’t as necessary.”
There is scope, potentially, for service providers and cloud brokers to step-in and relieve some of pressure on enterprises by rolling out their own IoT cost-optimisation services, said Rogers.
“They could offer enterprises access to all of the different cloud provider IoT platforms, and recommend which ones are best value for how those organisations plan to use it, and automatically move their workloads between them if their needs change.”
“A lot of companies, like Rackspace, are offering third-party management of cloud providers, and they are naturally in a good place to offer such services because they have the expertise to do so,” he said. “I wouldn’t be surprised if one day they start to expand more into IoT services.”