Operator edge computing initiatives
Since we last reported on operators' edge computing activities, a number have progressed to include work on integrating edge computing into 5G networks, now that the standards process has advanced further.
AT&T
AT&T says it is embracing edge computing for both its fixed and mobile networks. The company has continued to expand its plans in this direction by building an edge computing test zone covering several miles in Palo Alto, California, under its AT&T Foundry initiative, and working with third-party developers, startups and others, in order to test use cases, including autonomous cars, AR/VR, and drones. It has also embarked on a program to convert thousands of central office locations into data centers.
More recently, the focus of this work has extended to include support for edge computing in a 5G network, and specifically for industrial IoT and enterprise use cases where AT&T wants to provide businesses with lower latency, access to high compute power, and network routing without needing on-premises hardware.
Deutsche Telekom
Deutsche Telekom (DT) believes that IoT is an important driver for MEC, with low latency being the major benefit. The operator believes that using edge computing to offload intelligence and processing power from devices into the distributed MEC cloud can help to achieve a better form factor and price in the device. DT was early in establishing laboratories dedicated to edge computing in Pittsburgh (US) Berlin (Germany), and Krakow (Poland).
After an internal study, DT decided to move much of its edge-computing-related activity into a separate business, called MobiledgeX, with the intention of creating a platform to run on top of the edge infrastructure and enable the creation of edge-based 5G services. MobiledgeX does not sell to other operators, but rather partners with them to help monetize their network assets. Among those it has already engaged with are SK Telecom, Vodafone, and Telefonica.
DT itself is also understood to be investing in new edge computing facilities in Germany using technology developed by MobiledgeX. In September 2018, the company announced a partnership with Aricent to launch a developer platform called Open Source Edge, with selected components expected to start rolling out in 2019.
Korea Telecom
Korea Telecom (KT) announced in May 2019 that it would be deploying an IT edge cloud to enhance the overall 5G service quality across the country. The deployment will support applications such as VR and streamed games. The edge cloud works by directly connecting smartphones to media content without the need to go through the central server, as the communications center or radio base station nearby functions as a main server, shortening the process from four steps to just two, and reducing the data transmission time by 10 seconds.
KT says it will gradually expand its deployment coverage for IT edge cloud into eight regions in the country, initially starting with deployments in Seoul and Busan, the two major metropolitan areas.
Orange
Orange says that as a convergent operator, edge computing fits well into its future service strategy, particularly in its European markets. It does not regard edge computing as a technological challenge having trialed the technology in network tests for connected vehicles with Ericsson and Citroen, employing a 4G radio with a pre-5G architecture. Rather, it sees edge computing as being more about the business case. Orange is active in ETSI MEC but says it is not wedded to any particular approach and is currently analyzing potential architectures. The operator envisages three main use cases for edge computing:
- enterprise services employing virtual computer premises equipment (CPE)
- specific industrial sites, initially using 4G
- B2B and retail around VR.
Orange expects the three areas will develop with different timelines and deployments, with industrial sites able to be worked on as limited deployments without requiring significant, large-scale investment.
SK Telecom
SK Telecom is commercializing what it claims is the world's first 5G-based mobile edge computing platform built on the ETSI MEC standards. The edge computing deployment minimizes latency by installing a small-scale data center at the 5G base station or router. The operator says this will enable the provision of innovative 5G services such as AR/VR, cloud gaming, autonomous driving and fleet management, and real-time live broadcasting. SK Telecom also plans to broaden the MEC-related ecosystem by opening its APIs for third-party MEC service providers.
Telefonica
Telefonica is exploring the future role of MEC as part of its global network virtualization program, UNICA. Telefonica believes edge computing can extend the benefits of UNICA by being deployed close to the devices that generate data, such as in a central office, and has a key role to play in the development of IoT and by supporting the low latency that services and applications will need, such as 8K video, VR/AR, streaming video games, tactile internet, robotics, or autonomous vehicles.
Verizon
Verizon says it already has the capability to employ edge computing on its network due to its extensive deployment of Cloud-RAN technology and service access points. These locations, which are connected to the operator's cloud network and run virtualized network functions, could support edge services with only a software upgrade, according to the operator.
In January 2019, Verizon said it had halved latency speeds on its 5G network in a trial using video transmission and facial recognition software, and running on its Intelligent Edge Network (iEN). The operator is also reported to be testing a cloud gaming service on its network and is in the process of developing tools that will allow developers to create services using the low-latency speeds available through edge computing.
Verizon says that most of its efforts in the edge computing space will initially focus on enterprises rather than the consumer space.
Vodafone
Vodafone has been working with Saguna to quantify the impact of MEC architectures on video quality. Here again latency has proved a key factor, with the operator reporting reduced waiting times and video stalls, and a significant overall improvement in quality of experience (QoE). It also emerged in the course of the trials that congestion in the radio access network (RAN) affects video QoE of remote content more than it affects MEC-hosted content.
