Owen Rogers (that’s a likeness of him in the picture above), IT industry analyst from 451 Research, is something of a martyr. This is a guy who spends most of his waking moments obsessing over the minutiae of cloud pricing. If we think of a series of concentric circles, inside the circle that represents IT, is a smaller circle relating to cloud computing. Inside this circle is another circle, undetectable to the human eye, that is the place where all human knowledge of cloud computing pricing exists. This is the world inhabited by the aforementioned Owen Rogers.

Now you’d be forgiven for thinking that someone who inhabits such a dark and forgotten part of the world would be something of a recluse, perpetually pouring over ancient tomes and with an air of eccentricity and forgetfulness. But while Rogers is admittedly pretty pasty (he hails from Wales, after all), he’s managed to avoid the pitfalls of his chosen vocation and is actually a thoroughly fun chap. Just goes to show, huh?

Anyway, Rogers latest work has focused on a specific part of that already small circle. In particular, he’s been looking at the economics of the Internet of Things. Now the Internet of Things (IoT for those in the know), relates to all those connected devices that our present, and future, will comprise of. You know the thing – your connected toothbrush, refrigerator and, somewhat worryingly, adult toy. For the first time, Rogers and his colleagues at 451 Research have analyzed the costs across the different IoT platforms and have leveraged a bunch of technologies themselves (machine learning! no less) to clarify some different scenarios and their associated costs.

On to the findings

At a high level, 451 found that Microsoft Azure is the cheapest at scale while Amazon Web Services (AWS) provides better value for smaller deployments. In a finding that will annoy those at the Googleplex, especially coming as it does only a week after they failed to make up much ground in Gartner’s public cloud magic quadrant, across the 10 million IoT scenarios analyzed, Google Cloud was never found to be exclusively the cheapest provider.

This report actually comes at a pretty important time. Enterprises would seem to be looking to IoT en masse as the new frontier, but the complexities around pricing are something of a decision-making barrier. That’s not to say that enterprises are primarily driven by price, however, they do want clarity on what things actually cost – the complexities of cloud pricing are, witnessed by the fact that Rogers manages to spend his every waking moment on the topic, sufficient to make that a difficult ask.

The finer details

451 Research found that Microsoft is the cheapest provider when the message size, number of devices and number of messages sent are all relatively high. However, this is dependent on the buyer predicting their needs in advance. Meanwhile, AWS is the least expensive IoT provider for those smaller deployments more typical in enterprises today. In deployments of less than 20,000 devices each sending an average of fewer than three messages per minute at under 6KB each, AWS is the cheapest provider in over three-quarters of scenarios tested. Rogers comments on these complexities, saying that:

In theory, it should be simple to predict and account for pricing but the hyperscalers obfuscate this with subtle distinctions. Even when a pricing model appears to be fairly simple, the impact of underlying nuances is often unclear. With this new report, both enterprises and the providers have the transparency required to make better business decisions

The report finds that the costs related to AWS IoT Core, Google IoT Core and Microsoft Azure IoT Hub are structured so differently that manual cost comparisons are almost impossible. Even where pricing models appear simple, subtle differences can have significant cost implications.

One hidden element that affects cost is that each provider has its own definition of what constitutes a message. For example, a “keep alive” message of just a few bytes is free on some IoT platforms but charged on others. In fact, some cloud providers round-up small messages such as this to the nearest kilobyte. This results in a 64-byte “ping” message being charged for 17 times the capacity. So much for doing the right thing by customers, huh?

Arise the cloud brokers?

In something of a return to a discussion we were all holding circa eight years ago about cloud pricing brokers, 451 Research believes there is an opportunity for companies selling cloud services to become broker-dealers and launch offerings that resolve and hedge away this complexity, allowing enterprises to easily purchase an offering that suits their business without the worry of paying over the odds.

Delving into the methodology

As can be expected from a studious bunch, 451 Research thought through the methodology of this report. They discovered that nine individual pricing parameters have a bearing on the cost of the hyperscaler IoT platforms. With so many permutations, the only approach was to price all possible combinations and understand the differences using machine learning techniques.

The analysts constructed a Python simulation that performed 10 million comparisons of US pricing. Each simulated scenario had a randomly selected configuration of the 9 price-impacting parameters.

The training data produced from the simulation was fed into an analytics platform, producing a Chi-square automatic interaction detection (CHAID) decision tree with, the aim being of finding which combinations of the parameters drove which provider (AWS, Google or Microsoft) to be the cheapest. The tree had a predictive strength of 96%.

MyPOV

Apart from being adamant that Rogers deserves some kind of medal for the work he’s done here, the findings of this report make for interesting reading. Both Microsoft and Amazon will be generally happy with their results, while the Google exec team will no doubt be reviewing their pricing approach. One thing is for sure, if we were to do the same analysis in six months time, the results would be very different.

Ben Kepes

Ben Kepes is a technology evangelist, an investor, a commentator and a business adviser. Ben covers the convergence of technology, mobile, ubiquity and agility, all enabled by the Cloud. His areas of interest extend to enterprise software, software integration, financial/accounting software, platforms and infrastructure as well as articulating technology simply for everyday users.

Leave a Reply