There’s bad news and good news, writes Shamik Mishra, the associate vice president of the innovation team at Aricent. By next year, according to industry insiders, the over-the-top (OTT) market could double in almost every aspect that impacts mobile operators. Network providers are already losing ground to the likes of Facebook and Google and are haemorrhaging revenues.
On a positive note, a number of forward looking operators are exploring technologies such as augmented reality (AR) and virtual reality (VR) to attract subscribers and secure new revenue streams. Immersive media such as AR and VR can enhance the subscribers’ quality of experience (QoE) and it allows operators to deliver rich content via 3D images and videos. What’s more, the AR and VR market could be worth US$108bn by 2021. That is very good news, but unless mobile operators have the technology in place to be at the heart of an ecosystem, operators could merely end up being the connectivity provider.
AR and VR dreams
Mobile edge computing (MEC) provides operators the opportunity to monetize technologies such as AR and VR with the most important asset in their arsenal, their low-latency access network infrastructure. It is their beachfront property. Thanks to MEC, applications and processes can run on mobile network infrastructure that is closer to the user – dramatically reducing delays and increasing the performance of devices such as AR and VR handsets.
If you are watching AR/VR content or even playing a game with your headset on, – you don’t want video lags or buffering. That wrecks QoE. The AR/VR application algorithms, data augmentation and the displays work at very low latency and require considerable battery power. As a result, most headsets can be bulky, awkward and can even heat up – which makes for an uncomfortable experience. Given that some headsets cost around US$1000 or more, that’s quite an expensive headache to experience.
One solution for that could be to offload the computing from the device to the cloud. However, if the computing is done on a public cloud server far away from the device, the speed of light constraints would increase the latency and the QoE would deteriorate. With MEC, the application could be launched and the computing for it could be processed on a powerful server nearby – rather than several thousand kilometres away on a public cloud. This lowers and even negates latency. The AR/VR headsets could become slimmer with less processing at the device and the batteries consume less power. The cost of headsets could become cheaper too making AR and VR devices a personal accessory – not a sparingly used geeky, expensive gadget.
Take it to the edge
So, why does MEC give mobile operators an edge? Operators have the low latency access networks and the local data centres close to the subscribers. Carriers know the location of their users and can therefore find the most appropriate data centre for the device. As a result, data can be transmitted at ultra-fast speeds. This way, the latency targets can be met and the computing can still be offloaded. Faster computational and processing speeds also enables operators to launch new services and strengthen existing ones.
One potential new service could be real time object recognition. By integrating artificial intelligence (AI) and machine learning (ML) algorithms, operators can enhance the edge. This can be used for pedestrian or obstacle detection in autonomous driving, intrusion detection in video surveillance and for image and video search. The MEC platform hosting such a service needs to be intelligent in order to provide such autonomous services. It also requires supporting event-driven architectures where computing can be offloaded to the edge on-demand. Modern serverless architectures could be a potential solution for AR/VR computing offload to the edge.
Edgy mobile videos
In terms of enhancing an existing service, operators can use MEC to dramatically improve mobile video QoE. Around the world nearly a billion people alone watch Netflix and YouTube on their handsets. Mobile video consumption is growing exponentially and operators need to stay on top of this. Unsurprisingly, buffering is hardly tolerated by subscribers and they tend to blame operators for poor QoE.
Currently, videos are stored on data centres and delivered over content delivery networks (CDNs) that in some cases can be hundreds or even thousands of miles away from the user. The delay in dispatching and delivering that data can cause delays and impact QoE. What’s more, content providers can incur high packet transfer costs.
Mobile operators can cache popular video or live content at the edge and save backhaul costs. This provides an opportunity for OTTs and telcos to partner and deliver content faster with minimum latency and lags – and with reduced packet transfer costs. It is a win-win-win. What this use case also highlights is the paramount importance of having an edge-based ecosystem in place where the operator can be at the heart of it.
A cloud first mind set
The Silicon Valley technology behemoths which have come to dominate the cloud have embraced developers and provided them with control. They enjoy easy and flexible access to the public cloud and have software tools and monitoring systems for their applications. Mobile operators can go one step further and place developers in the driving seat within a thriving ecosystem. To do this, mobile operators must adopt a cloud-first mind set.
Network providers have the capability to create a software platform that is developer centric to help them make the transition from the public cloud to the edge seamless. To help developers even further, a cloud-first approach needs to transcend across the operator’s technology stack for edge computing with cloud native, serverless, platform-as-a-service and content delivery software systems.
In addition, most application developers have come to rely on the application programme interface (API) economy to access a variety of tools and skills. Mobile operators can create a platform where developers can access intelligence like face recognition, object recognition etc. for immersive media through APIs. To achieve this, operators have to embrace AI and ML technologies and provide developers with consumable capabilities such as learning engines and micro-services.
Show me the money
Ultimately, deploying MEC is all about securing new revenue streams and maintaining a healthy bottom line in the face of an ever-growing OTT threat. Operators must think imaginatively on how they would monetize and bill their users for this service. Traditional billing methods could be replaced with a pay-as-you-consume model. MEC is not just revolutionary, it is the natural evolutionary next step for operators and last but not least, the developers.
