The second fundamental change impacting third-wave product design is a new architecture. Given the ingredients of cloud technology, ubiquitous connectivity and sensing, a device-to-cloud architecture would seem right. Just capture the sensor data, upload to the cloud, and analyze. However, a device-to-cloud architecture (often called a cloud-oriented architecture) is the exception today, not the norm.
A distributed device-to-cloud architecture provides the flexibility required by third wave applications — the ability to distribute application logic across unit scale end-points, local scale edge-compute systems, and planetary scale cloud data centers. Unit, local, and planetary scale are the three tiers of compute scale.
Unit scale refers to systems that are constrained in terms of computing (thinking), networking (connecting), and power. Constrained resources mean a system has limited capabilities. Such systems tend to be designed only for a specific function such as data collection. Constrained networking means the system is likely to utilize Low Power and Lossy Networking technology, made up of many embedded devices with limited bandwidth and processing capability. Constrained power means the system uses batteries or relies on energy harvesting. The amount of power available on a constrained device is limited by design. For example, a wireless sensor network may employ unit scale systems for devices that must collect readings in the field for five years before running out of battery power.
Local scale refers to systems that have operating systems and general-purpose computing capabilities to run a wider range of applications. These systems are also known as edge computing systems because their location in the architecture is closer to the edge of a
traditional LAN or WAN. Gateways, laptops, smartphones, and appliances tend to fall into this category. Local scale systems are necessary in cases where access to the cloud is unreliable or there are real-time application requirements.
Planetary scale are cloud compute systems. This is not just a single web server on the public Internet. Planetary scale refers to the capabilities made available by parallel computing infrastructure and elastic on-demand capacity.
Choosing the right distributed architecture depends on the needs of your application. Cloud offers scale and parallel computing, edge provides real-time and general-purpose computing, and unit scale end-points offer highly power efficient, yet constrained computing capabilities.
The IoT is ultimately not a single thing – it is a series of vertical market solutions that will emerge and grow at different rates. There are very few commonalities across these various “Internets.” Many technology companies have looked for horizontal building blocks across use cases and verticals, but thus far those seem to be limited to connectivity and provisioning, and some aspects of data warehousing and security.
While the first and second waves of computing used these horizontal building blocks, the building blocks for the third wave look more like a haystack.
- At the bottom of the haystack are the sensors, actuators and equipment boards. There are a huge variety of these components, and how they are assembled and used vary dramatically by industry and use case.
- At the top of the haystack are the many applications and associated analytics capabilities of various IoT applications. Although there are a few conceptual commonalities – e.g., predictive maintenance for consumer cars versus predictive maintenance for automated factories versus “predictive maintenance” for consumer healthcare – when it comes to actually designing the applications and algorithms, there is little that can be leveraged across verticals.
- In the middle of the haystack are the horizontal building blocks, consisting of data storage and warehousing, device provisioning, device management and patching, security and compliance, and connectivity. Even these elements are not truly horizontal because things such as security requirements differ by vertical, and there are industry preferences for certain forms of connectivity over others.
The upshot is that even though it is tempting to think of the IoT as a set of horizontal capabilities connecting the physical and digital worlds, the reality is that it is a series of vertical use cases with incredible diversity.
Because of that diversity, third-wave success requires a stronger emphasis on building end-to-end systems. A complete understanding of the entire system and its various use cases is critical when products traverse multiple technologies, from sensor to gateway to cloud, and includes multiple user roles such as end users, field technicians and customer support, all with multiple touch points. Today’s emphasis on user experience and solutions that “just work” leave little room for players that do not keep track of all the moving parts and dependencies.
Come back later this week for more insights on New Software Skills and Security (Thu Feb 26) and New Market Dynamics (Fri Feb 27) that will shape the third wave of computing. For more perspective on IoT product engineering, please click here to access Altran’s full white paper titled, “Never Mind the IoT, Here Comes the Third Wave.”
Editor’s Note: This is the third in a five-part series on #IoT #Engineering trends that will shape how companies design, build, and bring products to market.
To learn more, please register
here to meet with our team at #MWC15.