Does IoT make sense for manufacturers?
There’s just one problem: all these ‘things’ require cloud connectivity. Cloud connectivity requires a data center somewhere on the Internet. It also requires a team of software experts to scale and maintain the cloud. As a manufacturer, if you sell thousands of cloud-enabled products for a one-time fee this year, are you willing to continue paying for the data centers and software maintenance to keep those products online 5 years from now? The math doesn’t add up, and many manufacturers are thinking twice about taking products directly to market.
There are some tough choices a product manager faces when designing a strategy to take on the Internet of Things. One option is to avoid the costs of maintaining a cloud by creating products that only work locally, but customers demand more. Another option is to build a custom cloud, but at the expense of months or years of development, the cost of hiring software experts, the fees of maintaining data centers every month, and the risk of quickly falling behind the technology curve.
Yet another option is to charge customers for hosting services, which transforms the relationship between customers and manufacturers. Hosting fees aren’t friendly to consumers, and it’s simply not competitive. Interview any of your friends and family, and you’ll quickly discover people don’t want to pay for a monthly fee for basic connectivity, but they would be willing to pay for powerful services delivered by having access to connected devices.
Fortunately, manufacturers don’t have to do everything themselves. Many IoT clouds already exist today, ready for manufacturers to snap their products into and go to market. Cloud services are separated from the physical protocols used in products, so the product’s physical protocol usually doesn’t matter.
ZigBee, Bluetooth, WiFi, Z-Wave, and other protocols ultimately converge to one protocol: the Internet Protocol. Of course, some protocols require an extra gateway, and others do not. Some protocols are more conducive to battery-operated devices, and others are designed for high power. The best option for a manufacturer is to simply choose a common, reliable protocol that will plug into a wider ecosystem and provide the best out-of-box experience for consumers.
The next step is to select a cloud to plug into. This is where manufacturers need to be careful. One thing is clear: most IoT clouds attempt to charge a per-device recurring service fee. And this brings us back to the problem of maintaining services for devices over the lifetime of the product. Do you really want to sell a product for a one-time charge, and be responsible for paying to keep it alive forever? Probably not.
Presto created by People Power Company, is one IoT cloud announced earlier in 2016 that offers free, secure, and scalable hosting for products. As a software company, People Power justifies a free product hosting model by selling white-label and people-centric services through massive service providers, which allows their cloud to be open and available for manufacturers (and provides new channels to market for manufacturers). With over 157 completely open APIs and open source reference code, manufacturers can connect devices with any physical protocol to the cloud and deliver cloud-enabled products and branded apps to market in a way that is compatible with traditional manufacturing business models.
Connected products form connections with cloud services from the device to the cloud, and not from the cloud to the device. By originating communications from the device, devices automatically break through NATs and firewalls that would normally block out external connections.
Products need to upload measurements and state information to the server periodically, and declare themselves ‘online’. Because data is the new oil in the Internet of Things, it’s generally considered a best practice to upload measurements more frequently when measurements are significantly changing, and slow down uploading measurements when there is nothing happening. This strikes a balance between maintaining a long battery life and delivering a responsive user experience.
Connected products sometimes need to receive commands. Near-instantaneous commands deliver the best user experience around the product, even if the user on the other side of the world. To deliver commands from a cloud to a device near-instantaneously, the device needs to form a persistent connection with the cloud. These connections can be formed with standard HTTP GET mechanisms, WebSockets, or even MQTT, as the cloud permits.
But each connection consumes a port on the server and servers only have a finite number of ports, which leads to a scalability problem from the cloud’s perspective. In this situation, expensive load balancers fail to add value because every port is already consumed and there is nothing to balance. Clever IoT clouds will split the load balancing functionality between the device and the cloud, allowing the device itself to choose the “best” server with which to connect, and enable the cloud services to scale gracefully.
Figure 1 shows the architecture of the infrastructure designed to process a large number of device API calls. The load balancing functionality is split between the device and the cloud, where approximately every 24 hours, the device will retrieve the address of the ‘best’ physical server to connect with. In Presto, this architecture has been used to handle millions of API calls per day in live IoT applications while minimizing cloud hosting costs.
During the week of February 8, 2016, manufacturers using Presto delivered about 30 million data posts from connected devices at an average response time of 500 ms. In parallel, the server processed 27.5 million analytic requests, or 46 analytic requests per second, adding value to the user experience around those products in the background of user’s lives.
The Internet of Things doesn’t stop at just cloud connectivity. Rules engines coordinate devices across different protocols and manufacturers. Trusted social networks enable caregiving and neighborhood watch programs. Community social networks drive behavior change in consumers. Real-time big data analytics enable connected products to do things manufacturers never imagined. Command centers enable service providers to deliver services to customers. And intelligently architected mobile apps enable consumers to have one app to control all the products around them, because nobody wants 20 different apps for 20 different devices.
As manufacturers move up the stack of the Internet of Things, they’re met by software companies who are working their way down the stack. It’s a huge stack of hardware and software combined, and nobody can build it all themselves. Innovative manufacturers and software companies must partner together to make the products not only work for real people, but work in a way that is compatible with manufacturers’ business models. That’s the only way to make the Internet of Things make sense.
About the author:
David Moss is President & CTO of the People Power Company – https://peoplepowerco.com
Related articles:
GE unveils cloud service for industrial IoT data & analytics
Electric Imp doubles down on IoT
GE IoT startup offers ‘new kind’ of energy company
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
