The internet of things is getting bigger. And it’s getting there fast. According to a recent infographic from Intel, the number of “things” in the Internet of Things grew from 2 billion in 2006 to 15 billion in 2015, and is expected to grow to a staggering 200 billion by 2020. That leads to a key question – can we build enough infrastructure to handle all those “things”?
It’s safe to say that there will be different requirements for IoT devices. Some will be self-contained and self-sufficient, not requiring any, or only infrequent connectivity. Think of oil well pumps, or maybe even your smart refrigerator. Then there will be other types of devices that may require a little more interaction and a little more connectivity. Say for example, onramp sensors and monitors that are connected, but don’t necessarily need instantaneous feedback from a centralized management system.
Then there are classes of devices, or collections of devices that create solutions, that need near real-time response. Or, they may be more autonomous in nature since their connectivity to centralized systems may be limited, but they still need to respond quickly to ensure public or personal safety, or to deliver an expected or contractual levels of service. These types of solutions are ideal candidates for “mobile-edge clouds.” Let’s take a look at a few examples.
Without a doubt, the explosion of video has placed staggering demands on the current network infrastructure model. 5G is poised to provide significant relief in terms of latency, but it can only solve part of the equation for new video-based applications. In its simplest form, optimizing network conditions in real-time for video acceleration is a requirement for pretty much all video-based applications. But much more sophisticated applications are placing higher demands not only on the transmission of video, but also on the processing and analysis of video content. Things like event triggering for movement, missing objects, crowding, etc. Or denaturing for security and privacy purposes. And facial recognition for threat analysis and response. All of these examples dictate that the processing of vast amounts of data be co-located or in very close proximity to the video source.
Almost everybody has seen self-driving cars, or at least seen or read about them on the news. There’s an ongoing debate as to how autonomous these vehicles must be, with most experts agreeing that they must be fully self-contained. But there’s another model for mobile automotive solutions – connected cars. In this scenario, intelligent cars share information about road conditions in real-time. With onboard visual systems they can detect and report things such as potholes, roadkill, or objects on the roadway. Or maybe traffic flow, or weather conditions. All of these situations require processing and sharing of data very quickly and reliably.
Immersive Games and User Experiences
Virtual and augmented reality are quickly becoming an essential element of most gaming experiences. And the games themselves no longer require (or expect) the player to be tethered to a high-performance desktop system. Real-time interaction between any number of participants requires that processing be co-located with the players to minimize latency. But gaming is not the only new user experience that is placing new demands on traditional infrastructure. Truly “mobile” services for cruise ships, trains and even airplanes, where connectivity is limited or maybe non-existent, require that processing be nearer to the user to not only deliver, but also tailor the user experience to specific buyer personas.
The need for agile infrastructure
These are just a few examples of how the Internet of Things is driving new demands for mobile edge services. Like IoT itself, the number of use cases for mobile edge clouds will skyrocket over the next couple of years. All these use cases will have one thing in common – the need for agile cloud infrastructure that can be distributed to hundreds if not thousands of endpoints. OpenStack is the obvious virtual infrastructure manager (VIM) of choice. But by itself, it doesn’t support a distributed cloud model. That’s where CPLANE NETWORKS comes into play. We have the industry’s most powerful and flexible platform for quickly and reliably building hyper-distributed clouds – Multi-Site Manager. Let us show you how you can build your mobile edge clouds for tomorrow’s applications – today!