Moore's Law in IoT will kill any hardware business

In 2015 we had the Raspberry Pi 2. A super computer the size of a credit card for $35.

In 2016 we got equally or more powerful competitors, half or a third the size, all under $10, e.g. Nano Pi Neo, Orange Pi Zero, Chip, and more:

We also saw less powerful alternatives around $5, like the Raspberry Pi Zero, Omega2 and others.

In 2017 we are seeing that developer boards now are getting cheap industrial siblings like the Raspberry Pi Compute Module 3 for $30 and the Chip Pro for $16:

It is just January 2017 so by December we probably see smaller, cheaper and better. However what will we see in 2018, 2019 and 2020? Moore's Law says every 18 months we will double speed, half price and size. So we will have a $1 super computer this size in the next two to three years:

With that form-factor and price it will be more expensive to solder a custom solution for a hair-dryer or electric toothbrush than to put a super computer into it.

In a world of 100's of super computers in your home and in your office there will be three major issues:

1. Security
2. Security
3. Security

No just kidding:

1. Security
2. Management
3. Monetization

How are we going to avoid an exploit from taking over our house, business and cities? Last year we saw how the open sourcing of Mirai, the botnet code that provoked the biggest distributed denial of service attack to date, allowed criminals to also combine it with a security flaw in Deutsche Telekom's broadband modems, i.e. Misfortune Cookie. The end result was that 900,000 broadband modems formed part of a targeted cyber attack that brought the Internet in Germany down for a day. What will happen if now each home owns 100's of super computers? How are we going to secure them?

We need to rethink our way we manage security and update software in devices. We need DevOps for Devices in which daily all types of security patches can be fixed in a fully automated and continuous integration tested way. If something does not work then we need to be able to roll-back to the previous working version.

We also need to rethink the economical model because on a $1 board, the margins are going to be almost zero. Making money on IoT hardware will become double as difficult, every 18 months.

So what is the solution?

IF YOU DON'T CARE ABOUT TECHNOLOGY, YOU CAN SKIP THE FOLLOWING PARAGRAPHS AND FORWARD THEM TO YOUR TECHNICAL TEAM.

We have been working on trying to solve these three problems and we open sourced what we have so far. Ubuntu Core is a special version of Ubuntu, specifically for devices. You can run apps on top called snaps. A snap is a new packaging format for applications on Linux. Every snap assumes third-party code is a criminal until proven innocent. Unlike our previous packaging (Debian packages) which assumed all code was innocent until proven guilty. So from a security perspective if your code will control for example a Bluetooth device and all of a sudden changes its mind [or has a malicious code injection!] then the operating system will not allow it. You can easily create snaps via Snapcraft in any language with any SDK or packaging tool. We don't believe in lock-in so the snap format, the snapcraft tool and the snap daemon are all open source and we have worked with other operating systems on making it work.

One of the key features of snaps is that developers can decide when their code is ready to go from nightly build (edge), to beta, to release candidate and finally to stable. This allows the developer to automate testing and very easily deliver security and other types of patches. You upload a new version to stable and all production devices will automatically start updating.

Ubuntu Core also uses snaps for the operating system and the kernel. All snaps are digitally signed so the device knows what exactly is running on top. If you make a mistake, then you can roll back transactionally. This means go back to the last working version and try again when you fixed your problem.

This solution will enable you to build completely automated DevOps for devices solutions.

BACK TO THE BUSINESS SIDE OF THE STORY

So with apps on devices how do you make money? You are probably thinking Canonical will run the app store and get all the money. Just like Google and Apple do. However think about what would happen if tomorrow there would be apps on an MRI scanner. Do you really think we have experts that can decide this app is safe and this app is not? We would soon approve apps that fry people's brains and stop other apps that cure cancer. Our lawyers suggested [STRONGLY] not to do this.

Our proposed solution is to let everybody else run their own app store. You know best what is important for your customers. What is good quality and what is not. So by running your own app store, you receive revenue from all the apps that get sold on top.

THIS SECTION IS ONLY FOR SMART AND BOLD PEOPLE

Have you ever wondered why Facebook paid so many billions for WhatsApp? Weren't they able to make a messaging app? Yes of course they are. But getting one billion people to download an app is the problem. If you have all your friends in WhatsApp and no friends in YetAnotherMessagingApp, then you aren't going to use the second one. This is called network effect and in the digital age we are seeing it everywhere.

We are going to show on mobile world congress that this network effect is really, really, really important for anybody thinking about making money from hardware. Our partners are going to open source hardware that previously was expensive. They will ask low-cost hardware manufacturers to make as many clones as possible. They will ask Amazon and others to drive out all the margin when reselling it. Hardware margins will be close to zero very soon.

So how do they plan on making money? Very easy by abstracting complexities and creating a network effect. If you make it super easy for an app developer to do something very complex like running a mobile base station or navigating a robot, then they can focus on making the real value add on top. It is this value add that will be uploaded to the partner's app store and which they can resell. So hardware clones will have an empty box. Hardware clones that collaborate and get their clone certified will have a vibrant app ecosystem. So our partners will make money both from royalty and certification of cloned hardware, as well as from the vibrant app ecosystem which gives them a revenue share each time an app is sold. They will not have to care about the dirty aspects of a hardware business like inventory, logistics, cashflow, and many more.

We are going towards a software defined everything economy. You have the open source tools to be part of this revolution. Our partners are already working on launching open source hardware products. What are you waiting for? Get in contact or make your first snap today.

----------------------

I believe that almost all LinkedIn connections are valuable. So happy for you to invite or follow me...