With the rise of the Internet of Things, there’s never been a more critical time for enterprise architects.
In a recent piece in AnalyticsWeek, Jelani Harper discusses how enterprise architecture — employing containers and microservices — helps smooth the path to IoT. “Microservices are especially well suited for the IoT because of the machine-to-machine capacity of the latter. In particular, numerous IoT or Industrial Internet deployments involve machine learning.” The ability to leverage IoT or Industrial IoT requires architectural thinking that looks at how the business and its customers can benefit from streaming data and accompanying analytics.
Along with leveraging microservices and containers to benefit the business, there is a need to determine how device networks will communicate, and where data will be processed. A new report issued by the US Government Accountability Office (GAO) observes there are four primary architectures associated with IoT, based on an Internet Architecture Board guidance document released in 2015. The GAO’s authors indicate four basic architecture models that can be applied to all IoT devices:
Device-to-device: “IoT devices within the same network that generally connect using wireless PAN protocols, such as Bluetooth and Zigbee, are device-to-device architectures,” the report illustrates. “
Device-to-cloud: In such architectures, “IoT devices connect directly to the cloud, typically using a long range communications network, such as cellular. For example, IoT-enabled vehicle monitoring devices (such as those provided by car insurance companies to drivers) collect data on the vehicle, such as distances and speeds driven, and acceleration and braking rates. These data are then transmitted to the cloud, analyzed in the cloud, and used by insurance companies to create tailored insurance rates based on the driving data.”
Device-to-gateway: “Device-to-gateway architectures transfer information from sensors to the cloud via a gateway device. The gateway collects the data and then communicates the data to the cloud through additional network connectivity, such as Wi-Fi or cellular connection.”
Cloud-to-cloud: “Cloud-to-cloud architecture, also known as back-end data sharing, enables third parties to access uploaded data from IoT devices. For example, smart buildings receiving data from smart thermostats and smart light bulbs can send the data to a cloud via Wi-Fi. The collected data are then aggregated in cloud 1, which may be owned by the building as the conduit, a user can set a smart thermostat to activate when the user’s car approaches their house (the conditional trigger in this example being the location sensor within the car).”
The rise of IoT is shifting software and network requirements within organizations moving forward with efforts, the GAO report also states. The report’s authors predict more emphasis on “analysis programs that can condense large volumes of IoT data into actionable information,” as well as “‘smart’ programs that can augment or replace a human operator. Aggregated data gathered from IoT devices can undergo sophisticated data analysis techniques, or analytics, to find patterns and to extract information and knowledge, enhancing decision-making.”
There will be changes in business structures as a result as well. “IoT software developments permitting automation may reduce the need for human operators in certain capacities,” particularly software that “relies on augmented intelligence and behavior to substitute for human judgement and actions, respectively. For example, IoT sensor data can be analyzed and then acted upon to reduce waste, energy costs, and the need for human intervention during industrial production.”
In addition, IoT may accelerate the evolution of business models toward software and data-driven companies. The most outstanding example of such a transformation is General Electric, which has “transformed its industrial line of business from the building and sale of light bulbs and appliances to a manufacturing and services line of business that not only builds complex industrial machines, but also provides an ongoing maintenance service offering based on the performance data gathered from the IoT technologies built into the machines it manufactures.”
New types of networks and protocols — such as 5G — are being developed to address evolving IoT hardware and software requirements, the GAO report adds. The 5G network, expected to be fully operational by 2020, was designed by members of the Next Generation Mobile Networks Alliance with 5G in mind, to enable faster communications between and with IoT devices.