It's increasingly possible for businesses to deploy connected devices without having to write any software.
Hardware modules and even complete devices which can be configured out-of-the-box to talk directly to IoT cloud platforms are appearing, accelerating the development of the IoT market.
In this two part series, two veterans of the IoT industry explore this theme. Starting us off this week is Pilgrim Beart, CEO of DevicePilot. Pilgrim has a history of founding connected companies, including Hive, sold to British Gas for $100m, and Antenova, selling billions of antenna systems and now in use in most mobile phones.
Is IoT evolving like the Web?
Let's look into the history of the Web as an analogy. In the early days of the Web, building a website required deep technical expertise to tangle with HTML, PHP coding and Apache web servers. But later, the arrival of off-the-shelf applications such as Wordpress, Google, and e-commerce apps made it possible for normal business people to get a website up and running in days.
IoT is undergoing a similar revolution: ten years ago, building an IoT proposition required deep technical skills, like designing and building hardware and its associated embedded software, understanding communications stacks and security, running databases - all inefficient, slow, and expensive.
In the past five years, the arrival of off-the-shelf vanilla IoT core platforms from AWS, Microsoft Azure, and Google has changed the market in two ways:
- Key IoT cloud components (edge connectivity, security, broker) are now available to all, with the advantages of SaaS: pay-as-you-go, at-scale quality, and pricing from Day 1
- Creating "watering hole" ecosystems where vendors and customers can easily find each other and inter-operate
Likewise, there's an increasing choice of IoT hardware - sensors and communications modules - and even of complete, finished IoT devices available for purchase.
Cloud-to-cloud integration is becoming quite straightforward, with no coding required. But building a complete IoT proposition, even from off-the-shelf IoT hardware and software stacks, still requires a significant amount of technical expertise and investment.
So how can we overcome this?
Configure and ship
Another useful historical analogy is the birth of Bluetooth. Bluetooth initially suffered from the fax machine problem - who cares if your phone has Bluetooth is there's no headset for it to talk to? Any company wanting to become a Bluetooth headset manufacturer had to pay a bill of materials (BOM) cost of around $5 and make a huge investment in software expertise, creating a big obstacle to market growth.
Cambridge Silicon Radio (CSR) cracked these problems to become the market leader in Bluetooth hardware. They solved the BOM cost problem by building a 2.4GHz radio onto the same silicon chip as a microprocessor - but how to address the software obstacle? Their solution was to ship a development kit which made it possible for anyone to design and ship a Bluetooth headset without writing any software at all.
A Tier 2/Tier 3 OEM in the far East could take the CSR reference design, run a simple configurator to set about 50 different parameters ("What does button 2 do?" / "Is echo cancellation available?") and then simply build and ship millions of headsets without having to write any software. No worries about the compatibility, security, etc etc. - the reference software takes care of all the gnarly details. This unlocked a Bluetooth market which now ships more than 4 billion units annually.
Could the same happen with IoT?
Say an ice cream business wants to track their ice cream through the supply chain to prove that it hasn't melted en route.
That's a pretty straightforward temperature-measuring application, so why should they build it themselves? They can just buy one of many temperature sensing devices, plug in a SIM, configure it to point to their IoT core platform of choice, and... ship it. No hardware expertise or coding required.