The “Internet of Things” is on the horizon. It is an expression that describes how the Internet will link traditional smart devices, along with a wide range of additional physical assets that allow these endpoints to generate and share data.
Nearly every product will have an Internet Protocol (IP) address and communication capability: not just networking and telecommunications devices, but also industrial equipment such as buildings, medical devices, test and measurement systems, construction equipment, and oil and gasoline machinery, as examples – that will link to other devices and services over the web.
This isn't so farfetched. In its September 2011 report “The Internet of Things Is Coming”, Gartner recommends that CIOs and IT leaders set aside time as early as “before mid-2012” – in other words, now – to develop a strategy for this scenario.
The trend of pervasive Internet is already gaining hold and is described today using a variety of terms in different industries: Machine-to-Machine (M2M) communications, intelligent device management, telematics, telehealth, and smart infrastructure are some examples. Gartner forecasts that there will be more than 30 billion permanently connected devices and more than 200 billion intermittently connected devices by 2020.
Opportunity beckons intelligent device makers. They must re-design products from fixed-function, disconnected devices to flexible, seamlessly connected systems. This can be done by understanding and adopting a software-centric approach to manufacturing and selling hardware.
In this new model, embedded software controls a devices' functionality or configuration. For example, a manufacturer might traditionally create a line of chips with two cores and a line of chips with four cores, costing X and Y respectively to manufacture.
Using embedded software – typically along with an embedded licensing and entitlement management solution – the manufacturer controls the proliferation of configurations by producing a single physical unit at a lower cost of Z, and the embedded software configuration controls the actual number of cores enabled.
At time of manufacture, some chips are loaded with a configuration for two cores and some are loaded with a configuration for four cores. The different chip configurations continue to be sold at different prices based on the varied configurations, but the manufacturing process and inventory is simplified and reduced. This lowers the chip's manufacturing costs by reducing the number of stock keeping units (SKUs), yet at the same time allows for the variability of chip models to satisfy diverse market needs.
Naturally, the design process for a device or component changes when using embedded software. Instead of manufacturing the full range of discrete device configurations, the embedded software is instrumented to correspond to the desired configurations. In the simplest model, at device startup time, the code reads the available configuration or licence rights corresponding to the number of cores. Based on the results, the chip determines whether to behave as the two-core model or the four-core model.
Determining what the different configurations describe is an important part of the design process, which differs based on the type of device. In some cases, configurations are a simple on-off switch for a discrete part of a component; in other, configurations describe a “value” in a range or collection of values (for example, device capacity – the number of ports enabled for a router or the resolution enabled by a camera). The two-core or four-core chip scenario described here is a borderline case that could be implemented using either interpretation.
In addition, the use of embedded software enables new value opportunities such as the ability to design a device that supports field upgrades. A field upgrade enables a lower-end system user to later upgrade to a higher-end system, without having to return the device to the manufacturer or dealer to be replaced with a new higher-end unit.
This is an important benefit of Internet-connected devices. Instead, the manufacturer designs the embedded licence rights to be updateable, providing a mechanism (such as an application that runs on the host system) that enables the customer to purchase an upgrade code that unlocks the additional tier of functionality for that device. In the previous example, an internet-connected version of the two-core chip configuration could be upgraded to the four-core chip configuration.
This type of embedded licensing model has been successfully used in many devices – mobile phones with unlockable GPS functionality, routers sold in tiers based on number of supported ports, and cameras with different signal-processing algorithms based on available licences.