By Rui Frazao, Vasona Networks CTO
Realizing the full promise of the internet of things (IoT) means creating an environment where tens of billions of devices can communicate in real-time, transforming industries. It’s been a long time coming, but application developers, enterprises and vertical industries are taking steps to make the IoT vision a reality. To fully realize the vision, they’ll need networks to support them. But are operators ready to deliver?
The mobile industry’s near-term and long-term IoT deployment needs can only be met with optimized network architectures. Operators evaluating this opportunity are often focused on the best way to meet low-bandwidth and low-latency connectivity. They know many of the required networks will require significant investments. To this end, the McKinsey Institute estimates that the total IoT market size will hit $3.7 billion in 2020, up from just $900 million in 2015. The revenue potential is accordingly large. Bain & Company expects that by 2020, annual revenues could exceed $450 billion for the IoT vendors selling the hardware, software and comprehensive solutions that will comprise the IoT ecosystem.
As the IoT market develops, there will be a huge range of use cases that need support. In particular, video-based IoT applications will demand specific network and technology approaches to meet unique requirements, including those that must leverage edge computing. These use cases will be low-latency, but high-bandwidth, with early examples including mission critical vehicle-to-vehicle communication and augmented reality. When operators design approaches to carrying this traffic, they will need to be able to understand the different video types traversing networks and assess which are most critical in the moment. The network edge will play a significant role in supporting this functionality and subsequent action.
Video-driven IoT use cases that demand edge solutions
Connected cars: As more connected cars hit the streets, sharing valuable data about the road ahead with other vehicles has the ability to save lives, as well as time and fuel. Auto manufacturers want infrastructure so they can share video feeds in real-time between vehicles. Of course, in an environment where the situation changes in the blink of an eye, real-time data exchange absolute requirement. To support split-second machine-to-machine decision making, there is no time for video to be transmitted via the network core and back out to vehicles. Instead, the traffic must be bridged at the network edge from one car to another in a location connected through neighboring stations at radio aggregation points.
Bridging traffic at the mobile network edge within the radio access network (RAN) is a new capability that network architects had not considered, until now. Previously, the concept was to push all types of traffic to the network core, close to internet connection points, where intelligence was built to take action on all types of traffic. This has been a successful approach since the majority of applications either download or upload information from the internet. But it introduces unnecessary delay for time-sensitive peer-to-peer traffic. This reinforces the need for a flexible, edge-based solution.
Augmented reality: Augmented reality in the enterprise is emerging as an early use case that is driving IoT video. These applications require head-mounted displays or mobile devices to hand over video to a local application, with data flowing back through the radio network. We’ll see this become increasingly common in the enterprise, such as verticals with distributed maintenance teams that deal regularly with complex repair scenarios. The success of the AR experience will depend on information being presented to the user in real-time.
A viable approach to meeting these stringent needs is using multi-access edge computing to locally break out AR traffic from traffic that does not have the same low latency and bandwidth requirements to ensure that it can be handled properly for an optimal experience.
From a privacy standpoint, the network must protect the privacy of individuals. This can be done by allowing only certified devices to participate in communications happening at the edge. Local breakout into a separate VPN secures the traffic and isolates it from the public internet.
Security: Surveillance and security initiatives can also benefit from the edge, especially to improve identification and response time to potential issues. For example, instead of a security camera transmitting continuous video through the core to a data center, the video stream can be processed at the edge via local breakout. If it captures a car sitting for an amount of time, pre-identified as an abnormal occurrence, the system could trigger an alert along with license plate and facial recognition. The video capture would transmit back to the core to notify the relevant parties.
For video-based IoT deployments, network knowledge matters
If operators thought they had a full plate keeping up with the time-sensitive needs of streaming video on networks, some may be wringing their hands thinking about how to address video-based IoT applications. But there is a clear path forward. It begins with a focus on network knowledge and acting in real time to properly match application performance requirements with network real-time capabilities.
In other words, it’s no longer enough for operators to see “video” on the network. They will need next-level and real-time analytics. They’ll need to understand which video sessions are mission-critical and low latency, and which are just low latency. And then, they will need to act on this information, determining whether to let the traffic flow normally via the network core or be delivered instead via the network edge.
Down the road, 5G is likely to provide more options for addressing an expanding set of video-based IoT use cases. But there’s no need to wait for 5G. In the meantime, operators that want to immediately pursue this market must begin meeting the need by creating edge capabilities in existing networks. As operators start taking steps, they will find the edge provides a perfect bridge to future architectures.