::: nBlog :::
While talking to our numerous industrial customers, ranging from heavy machinery suppliers to workforce tracking and navigation solution providers, I’ve noticed that my message is evolving into a quite standard form, even to businesses in very different segments and phases. In order to help them to create new products, to optimize existing ones or to just boost overall efficiency with IoT and spimes, I’ve gathered a few fundamentals to serve as a kind of Ten Commandments of IoT.
1. Be paranoid on security.
Very soon, most of your intellectual property lies in algorithms and digital recipes creating and controlling physical things. You must treat security as a continuous process and have defenses at all layers. Also, strongly authenticate, identify and audit all things and transactions. Some people say blockchain is the divine answer here, but there are earthlier and working methods available, too. Furthermore, make sure you give your customers control over their data – most will anyway give you access in exchange of better service. It’s all about trust.
2. Ensure scalability.
This might be the most important commandment, especially for customers planning to process data from millions of end customers. It is not only technical or architectural challenge, but also very much financial. When the number of your data points and algorithms soar, make sure you can sustain your business model with your key providers.
3. Design your product as global.
This requires that your provider has global reach what comes to networks and data centers. You don’t want to end up building separate backends in different continents or worse, countries.
4. Build and develop multiple watchdogs and failover mechanisms.
In global communications, life, universe and everything sometimes malfunctions. In a billion device environment, your small programming glitch easily concerns hundreds of thousands of units.
5. Prepare for ever faster release cycles.
Your product will inevitably update in a faster release cycle than before. Make sure your provider is in sync with this and won’t become a bottleneck if and when you need to roll out in 20 countries simultaneously.
6. Build for always-on operation.
In traditional IT and telecoms, service windows and outages are still commonplace. This is more of a process than technical problem. If you’re serious with your project, your IoT provider shall guarantee 24x7x365 operations using resilient infrastructure. When you provide e.g. heart rate sensors for hospitals or sensor-enabled call center for law enforcement, outages or service windows are just out of question.
7. Store the raw data.
This is one of my oldest lessons from my military times – nested algorithms are prone to systemic errors, which can only be ironed out by getting back to original data – if it still exists. Furthermore, the seemingly unimportant data can quickly become useful when new correlations are detected with e.g. neural networks and AI.
8. Build telemetry to everything.
Your customers will find unimaginable ways of utilizing your products, and you need to learn from that. Therefore, make sure your telemetry and debugging interfaces evolve with your product.
9. Maintain and grow your intellectual property.
During a fast-paced IoT product creation involving multiple stakeholders, lax IPR management will bite back seriously when the product gets to a billion customers. The always nice offshoring company might suddenly get greedy.
10. Build in sustainability.
When selecting your core suppliers, make sure their strategy is linked to yours what comes to your product roadmap. A service designed for the latest entertainment app might disappear in a year, potentially leaving your product stranded, no matter how large the supplier. Think of Google Nest.
By following these guidelines you’ll be prepared to transform your products into intelligent services and thus grow your business sustainably. Much like in the military, better and real time information gives you the edge. Watch our blog for the luring Fire-and-Forget attitude; either your sales guy (or AI) must be alert the moment your customer starts thinking about upgrading your product, even when this happens years after the initial purchase.