For manufacturers of connected products, Internet of Things (IoT) scalability means exploring a vast new wilderness. The IoT is brand new, and no one knows exactly how thousands or millions or billions of connected products will work — or not — as their numbers and interconnections multiply.
Scaling an IoT Product
The first place to think about IoT scaling is in the connected product itself, and with the embedded components that deliver IoT connectivity. The two main issues for IoT product manufacturers are intelligence and cost, which tend to move in opposite directions: More intelligence adds to the development costs, but less intelligence can compromise design goals.
The big question for balancing device intelligence and cost is: How dumb can your connected “thing” be? That’s a question that each manufacturer needs to ponder, because there’s no easy or one-size-fits-all answer. In part, it depends on what, exactly, needs to scale in any given IoT product. Regardless of the design and scaling needs, however, a few principles hold true for IoT product scalability, including:
- It’s important to be network agnostic. IoT products can use many underlying networks for connectivity to an IoT cloud, including Wi-Fi, Bluetooth, cellular, Zigbee, and others. It’s unlikely that any one network type will “win” the IoT, and even less likely that a manufacturer could guess which networks to bet on. Better, then, to cover all the bets by working with any network type.
- As the IoT evolves, embedded chips will see more processors designed specifically for the IoT. While first-generation IoT chips were too small to do what was needed, especially for security, chip vendors are incorporating more security and low-power features into lower-cost offerings. Trends toward more full-featured and capable IoT chips will aid the scalability options for the IoT products they power.
- Overcoming power and battery challenges are key to scalability. When communications modules embedded into IoT products transmit data, they tend to use a lot of power. Embedded IoT chips must be designed to support the ability to turn off parts of their capabilities when not actively transmitting, or to turn on full power only when transmitting. In addition, designers can use innovative compression technologies to reduce battery drain and use power more efficiently.
- Manufacturers need to realize that they probably lack the full range of expertise needed to design, build, deploy and maintain IoT connected products. They must become familiar with complex networking and cloud computing concepts; design hardware and write software code for reliable communications over multiple network types, including public data networks; enable end-to-end networking communications and security; and be able to design modern user interfaces for end-user management of IoT connections and devices. Manufacturers of traditional products have not had to deal with these kinds of technology issues before, and it takes considerable time and money to hire or build the expertise in-house.
- IoT products differ fundamentally from traditional products in that they are not static, they are not simply built and shipped and that’s that. Through capabilities such as over-the-air (OTA) communications, IoT products can be updated, connect to new services and even add product features in the field, years or even decades after they are deployed. This flexibility represents another kind of scalability, and one that traditional manufacturers need to understand.
OTA communications bring up another important aspect of IoT product scalability: future-proofing the product. Say a manufacturer designs an IoT product able to do five things today. But what if in a couple of years, there are 10 more things the installed product could do that could be delivered via OTA communications — but that the product hardware doesn’t support?
To ensure true scalability of the IoT product, manufacturers must think ahead about their designs to accommodate possible, as-yet-unknown future additions or expansions of functionality. Future-proofing IoT product designs might involve physically separating device provisioning, networking communications and security from the application used to control the device; creating an interface to the application processor that’s agnostic to the type of network and security protocols used; or providing improved versions of software with different networking options.
Scaling an IoT Business
In addition to the IoT product, IoT scaling must also be managed at the business level, as well.
Here are some things that companies will need to contemplate from a business perspective as they scale their IoT product deployments:
- Product support. Supporting a connected product in the IoT presents completely different challenges than supporting a static product that is sold and installed, then perhaps repaired if something goes wrong. Manufacturers of IoT products need to add software support, maintenance that is handled remotely or preemptively and new kinds of warranties.
- Customer support. People have different expectations from an IoT product than from a traditional product, and their concerns span things like networking, cloud and mobile app issues — all of which are new to traditional manufacturers.
- Software upgrades and bug fixes. This is an area that few traditional manufacturers had to face before the emergence of the IoT.
- Manufacturing forecasting. The longevity of a connected version of a product might be considerably greater than the longevity of a fixed product, which can create serious forecasting challenges for manufacturers.
- Pricing. Moving from a one-time product sale to an ongoing services-based model means rethinking everything about how an IoT product is priced.
- Version control. How many versions or iterations of a product will be supported, and for how long?
- Geographic expansion. Connected products, by definition, lack the same level of geographic identity associated with static products. Manufacturers might have to open or close regional offices to better support their IoT business.
- Staffing. Manufacturers of IoT products will need to hire employees with different skill sets than they needed for manufacturing traditional products, including things like network security and mobile user experience design. In fact, there will likely emerge an entirely new role with manufacturing companies — IoT engineer — to manage and coordinate the new capabilities and processes required to manufacture great IoT products.
The Role of an IoT Platform
In IoT scalability, technology choices matter — and how the technology is used matters, too. A specialized IoT platform should follow the 80/20 rule: delivering 80 percent of the IoT capabilities important for scaling at all levels, from connected product to cloud to mobile app to the data generated by IoT devices.
An IoT platform should provide a manufacturer with a complete IoT package: end-to-end coordination, security, reliability and scalability. Everything should work together seamlessly, and the IoT platform should be flexible enough to evolve and scale into unknown territory.
Oliver Cockcroft is product architect at Ayla Networks.