Security always should be top of mind when launching any IoT initiative, and we posted a blog in late 2016 specifically about this topic. What about practices outside of security? IoT is so new and moving at such a rapid pace that knowing the right approach is crucial to your success. One need only look at the graveyard of failed dot com companies to see the risk of jumping into a new tech sector without a solid strategy or understanding. Here are some tips to get you started thinking:
Get on the same page, every time
IoT-enabled devices theoretically are capable of anything, and that’s the danger. Even before discussing the technological aspects of deployment, you and your customers need to be very clear on the objectives and parameters. Not just the number of devices, but what are the devices doing (or not doing) and how will the collected data be used. The more specific the contract, the better for every stakeholder, especially if the goal is to have the devices enhance/generate revenue in some form. In addition, always ensure there are clear and legal lines of distinction as to data ownership and responsibility. If a customer violates privacy laws while employing your solution, for example, your company could be liable as well.
Find the right platform, or invest in shovels
New prototyping platforms have made low-volume IoT solutions simple and cost effective versus traditional technologies. But such solutions are not easily interchangeable, so you need to know your business model and type/number of devices before committing resources to a platform—software updates only go so far. Choosing the wrong platform can place you at a competitive disadvantage and, if the devices are embedded (such as with a roadway sensor), you must employ expensive human labor to retrieve them—or abandon them altogether.
K.I.S.S. temptation goodbye
Research has shown most people will fill their homes to the square footage available, regardless of size. This same psychology is rampant in IoT device development. Advances in technology have enabled a lot of software to fit in a relatively small hardware footprint, tempting designers to create top-heavy applications with unnecessary functionality. It’s like ‘deploying’ a Porsche to drive exclusively up and down the driveway.
The truth is most IoT-enabled devices do simple things and do them well. Having a home security system that also can tell how often the dog wags its tail benefits no one, not even the dog. Always have the objective of the deployment top of mind, and always employ the Keep It Simple, Stupid (no insult intended) or K.I.S.S principle. More technology means more complexity, and that can result in everything from security vulnerabilities to inconsistent reliability.
Nothing is set in stone, so don’t deploy that way
This may sound contrary to what’s discussed above, but it actually is part and parcel of the whole. Yes, finding the right platform and keeping things simple are mission-critical initiatives, but there is a twist when it comes to design philosophy. Traditionally, embedded controllers were designed with a ‘set it and forget it’ mentality, and with good reason. Most were deployed in isolation, without being connected to other systems or networks. Not surprisingly, this same approach does not work for IoT. The highly interconnected nature of IoT means products must be designed to be agile, with the ability to remain ‘future proof’ for as long as possible.
Your devices must pass every test before graduating
The nature of analog devices meant testing seldom involved a human, or any situation beyond the device’s basic mission in life (such as detecting a fire). IoT devices are a different animal altogether. They are so interconnected with the world around them, and can do so many diverse things, that you have to test for every conceivable scenario, including ones involving human interaction. Will you be able to predict every possible challenge? No, but proper testing will put the odds distinctly in your favor.
The good news is that many IoT devices can ‘heal’ and ‘think’ for themselves. Therefore, if your testing is done right and worst-case scenarios are discovered, the devices themselves can rectify issues on their own, even security breaches.
To learn more on what your devices can or cannot do, visit Neo.