June 21st 2020
Adrian Goins is Director of Community and Evangelism at Rancher Labs, which he joined in 2017.
The arrival of 5G wireless technology is often touted for the orders-of-magnitude boost in bandwidth it will bring to smartphone users and how it will make their lives wonderful, but the more immediate—and foreseeable—the impact will be on the Internet of Things (IoT).
Self-driving cars, power grids, remote medicine, and surveillance systems stand to gain significantly from the fifth generation of wireless technologies. And what will make it work is Kubernetes?
The Internet of Things, which has grown substantially over the last decade, is familiar to a lot of people through smart home devices like thermostats, alarm systems, and Alexa, but it primarily encompasses networks of sensors and devices feeding into public services such as traffic control, weather systems and health care.
IoT devices generate massive amounts of data, but they also tend to be low-powered with limited computing capability. Getting them to the point where they can sort information and pass along the most pertinent data to aid in decision-making depends on getting processing power and management closer to those devices.
While 5G can enable the necessary connections, a containerized orchestration system like Kubernetes is necessary to deliver the right information to the right place in as close to real-time as possible.
The Hand of Kubernetes in the IoT Glove
The Internet of Things already produces an astounding amount of data, which will grow exponentially as 5G networks are deployed. Managing that data and putting it to use is the challenge.
Traffic cameras, weather sensors, power meters and the like generate information that, combined with data from other cameras and sensors in a smart city environment, could be too much to process at a central location, especially if you’re looking for the devices to react to events as they occur.
A more efficient method would be to focus on smaller snippets from those sources based on a single goal and manage everything via automated processes.
In a downtown area, for example, traffic-light cameras and road sensors could keep track of the rate of cars passing a certain point on one block. If the number of passing cars reaches a threshold, that information could be parsed and sent one level higher up the network chain.
A decision can be made there, for instance, to reroute traffic or change the timing of the lights. A Kubernetes orchestration system enables that, with distributed processing power housed in software containers, which are automated and repeatable software units that always perform the same way.
Coupling containers with sensors, and managing their interactions, is the key to putting processing power right up at the network edge.
Those smart city cameras might also be doing other things like looking for people running red lights or making illegal turns or even observing foot traffic.
They could do all that while also monitoring congestion because Kubernetes breaks down operations of the devices on a network into discrete units of functionality. Each function can be managed individually in a service-oriented architecture, or SOA.
If you were looking at an SOA in the form of a management-type of the flowchart, each container would be responsible for certain key tasks. And then, depending on the data gathered or processed by containers at the network edge, they would report key bits of information to other containers higher up the chain.
Those containers would perform new functions with the collected data and then send their own reports. Eventually, a container programmed with more managerial functions would collect enough data to be able to make a reliable decision or perform an action.
As evidenced in the example, Kubernetes represents the next step in the evolution of managing computerized workloads. The orchestration tools allow for much better data management and decision making that is able to feed pertinent information to artificial intelligence or other analytics systems, effectively adding decision-making processing structures around inputs.
Since debuting in 2015, Kubernetes has become the de-facto standard for cloud container orchestration and is being applied to other environments as well. It doesn’t do the analytics or processing involved in IoT or cloud systems, but it makes them more efficient, scalable, and manageable.
Inside the Internet of Things
The Internet of Things comprises physical and digital devices connected to a network (and possibly to the Internet) that can transmit and receive data on their own, without human intervention.
The advent of 5G promises to make a lot of the futuristic visions of IoT a reality. From a consumer point of view, nearly unlimited bandwidth and signals that don’t stall or end prematurely will make it easier to download and stream movies, participate in augmented or virtual reality games, and interact with home appliances, health monitors and personal devices like smart drinking cups.
And those virtual meetings that people have had to get used to recently will also work a lot better.
Indeed, 5G will greatly increase wireless speeds, eventually delivering about 10 Gigabits/sec to smartphones with stronger signals, reduced latency, and better coverage in both remote areas and densely packed urban settings. But the real consumer benefits likely haven’t been dreamed of yet.
IoT likewise will have unforeseen activities resulting from 5G and other technology developments, but there are also a few plans in place for what’s next, at least for some sectors.
In the field of medicine, solutions such as remote health care, patient monitoring, and even remote robotically enabled surgery are on the table. In industry, the possibilities brought by machine-to-machine communication and analytics, along with a generous dose of robotics, are among the steps that will boost productivity and performance.
In transportation, IoT will essentially pave the smart roads, signals, and traffic management systems necessary for self-driving cars to actually work reliably and safely. Public transportation, in the form of flexible, responsive bus schedules or more efficient airports, will also benefit. And surveillance systems—whether seen as a great enhancement to security or an egregious invasion of privacy—will become much more powerful.
Telecommunications companies have been gradually launching 5G networks around the country, with services appearing in some cities in 2018 and 2019, and more extensive rollouts planned for 2020. Eventually, 5G will provide faster speeds, instantaneous connections, and the ability to network-connected devices almost everywhere.
And that’s where the ability to manage those devices—and more importantly, the data they produce—becomes critical.
The bigger and more powerful the Internet of Things becomes, the greater the need for precise management. In addition to managing containers and clusters of containers in a distributed system, Kubernetes is also capable of scaling with larger deployments.
Say, for example, a city has developed a very efficient traffic management system with a wide array of sensors and effective analytics software, connected on a 5G network and managed by Kubernetes.
If someone wanted to pass that system on to 100 other metropolitan areas, they would need to produce a large supply of identical clones that will work in any other city the same way as they work in the home city.
Telecom companies themselves are beginning to move toward using Kubernetes for their 5G networks, looking to take a software-defined approach wherever possible.
The 5G revolution is in full swing, and the Internet of Things keeps adding billions of new sensors and devices. But it will take an orchestration system purposedly designed for those unique environments to develop and unlock the true potential of the innovative applications that will eventually thrive in them.