How Mobile Edge Computing is Enabling 5G

5G, the next generation wireless network, is purported to revolutionize the way we interact with technology. In fact, the CTIA predicts that for the United States, $275 billion will be invested into building out 5G infrastructure, creating 3 million new jobs and generating $500 billion in economic growth. The arrival of 5G signifies the start of a fully connected, intelligent world, one in which new opportunities will arise for telecommunication providers (telco), enterprises, and end consumers alike.

5G Standards

The 3rd Generation Partnership Project (3GPP), responsible for developing standardized protocols in networking, has set the following benchmarks for 5G performance (see image). These improvements, as denoted by the ITU Radio Communication Sector (ITU-R), can be categorized into three main areas of application, each of which increase the scalability, efficiency, or reliability of existing 4G solutions.

Performance Metrics: ETSI

1. Enhanced Mobile Broadband (eMBB) – An increase in network throughput, capacity, and connectivity, benefitting high traffic areas in particular.

2. Ultra-Reliable Low-Latency Communication (URLLC) – Guaranteed < 1ms latency and high uptime for applications with such requirements.

3. Massive Machine-Type Communication (mMTC) – Connection between large amounts of low data-usage devices (e.g. sensing, metering, monitoring).

These feature sets enable a variety of previously impractical use cases across industries: cloud gaming, on-the-go video streaming, factory automation, autonomous vehicles, AR/VR, smart venues, and remote surgery, to name a few. However, to fully take advantage of what 5G has to offer, one other technology must come into play.

Mobile Edge Computing

Mobile edge computing (MEC) refers to extending cloud computing capability to the edge of a network. In this context, “edge” simply refers to the geographic location closest to where data is actively being processed (i.e., close to the consumer). Typical cloud computing architecture employs a few large server centers, each servicing large regions, through which all data is routed for processing—MEC, in contrast, decentralizes the network, introducing smaller nodes servicing more local regions. By bringing computing power even closer to the consumer, network latency is decreased, and computational offloading from central servers reduces overall network congestion. MEC can act as an enabler for many use cases when integrated with the high-speed, wide-reaching connection that 5G networks provide. To cite one example, autonomous vehicles would become more secure, equipped with the ability to process and classify surroundings in real-time. Clearly, a value proposition exists, but who stands to gain from it?

Emerging Ecosystems

Given the 5G era has only just begun, it is too early to tell how the ecosystem will develop, but there are a few key stakeholders:

1. Telcos – Telecommunication service providers such as Verizon, SK Telecom, China Mobile

2. Cloud Providers – Cloud computing service providers (Amazon Web Services, Google Cloud Platform, Microsoft Azure, etc.)

3. Business – Companies looking to leverage capabilities of 5G/MEC

As telcos start to expand their 5G networks, they have a variety of options for monetization, given their first-mover status. One possibility is that they work with cloud providers to consolidate 5G and MEC capabilities into a singular platform that utilizes tools developers will already be familiar with – effectively acting as edge hosts. This is already being explored through AWS Wavelength, a partnership between Verizon and Amazon that situates compute and storage devices directly at Verizon’s 5G edge. Amazon’s Matt Garman and Verizon’s Tami Erwin have shared promising initial results:

“We’re already seeing customers innovating in industries as varied as healthcare, entertainment, manufacturing, and sports. For example, Avesha ... is working with several hospitals to test machine learning inference at the edge using Wavelength and Verizon 5G to assist doctors in real-time detection and identification of benign and potentially malignant polyps … machine learning inference on streamed video and imagery data [requires] having a dedicated server with a GPU in the procedure room. If the system is not fast enough to keep pace with the endoscopic procedure, it could miss identifying polyps. But now, the Avesha platform is able to connect the procedure room with the inference models on high-performance GPUs at the cloud edge and the cloud backend that continuously updates the models, resulting in low latency performance with very high accuracy."

Another possible option would be for telcos to provide edge solutions themselves, forgoing partnerships with cloud providers. This is unlikely simply because telcos may lack the requisite domain knowledge to build and scale their own cloud infrastructure to a competitive level.

The prior two ideas assume that telcos will take a hands-on approach in incorporating MEC into their service models. However, it is not hard to envision a world in which telcos instead choose to focus on their area of expertise, leaving the onus on businesses to provide their own MEC infrastructure and real estate. The telco then provides a network orchestration service – some capabilities may include network provisioning (see 5G slicing), monitoring, and security tools. In this model, businesses would likely receive MEC capability from cloud providers (for example, AWS Outposts provides this functionality), install the hardware on-site, then tailor the telco’s service to their specific needs.

Suppose, for example, that a sports event is being held at a “smart” venue, where fans are guided to their seats by AR, live video feeds are played back on devices, and concessions can be ordered through mobile, directly to the location with the shortest line. Utilizing a telco’s orchestration service, the event planner can request a unique slice (an on-demand virtual portion of the telco’s network with resources tailored to meet specific service-level agreements), and will be able to monitor the network speed and latency to ensure proper functionality. Should these performance indicators fall below QoS, self-diagnostics and/or service employees may be available to help. Such a solution would be invaluable towards enabling next generation experiences, increasing interactivity and immersion for the end consumer.

Outlook

Being a relatively new network architecture itself, mobile edge computing has yet to become adopted as an industry standard, in part due to the shortcomings of current network technologies. As it stands, 4G lacks the requisite bandwidth to enable the aforementioned use cases on a large scale basis. 5G will thus prove to be a key driver in the growth and adoption of MEC architecture, and only the tandem development of these two standards will provide the coveted combination of high speeds, low latency, and increased network capacity on which a new generation of applications can be run. With companies like Verizon and AWS already exploring this opportunity, it seems likely that such an ecosystem will come to fruition in the near future. Then, it will be up to developers to design software that is more scalable than ever before, platform-agnostic, and capable of supporting distributed architectures such as MEC.