Edge computing strategies of the major vendors

Nokia

Nokia believes that the growth of edge computing plays to its strengths in cloud and packet core networks. The vendor's converged edge cloud solution supports Central Office Re-architected as a Datacenter (CORD) and MEC architectures, and forms part of Nokia's Future X vision for a converged digital ecosystem, alongside features such as the smart network fabric and universal adaptive core network. The vendor claims its end-to-end edge cloud solution surpasses best-of-breed alternatives.

Nokia's MEC platform includes:

  • MEC Access Edge computing platform, a software-based edge cloud solution designed to deliver flexibility, scalability, and efficiency to multiple base station sites using existing interfaces and Ethernet-based fronthaul. Nokia says the platform enables new possibilities to serve the operator's radio network and to coexist with other virtualized network functions
  • Zone Applications, a MEC platform for service providers and venue owners of locations such as stadia and transport hubs, which supports applications for visitors and staff, including live video feeds, instant replays, AR, video communications, and surveillance
  • Virtualized MEC (vMEC), a software-based solution that runs on commercial off-the-shelf
  • (COTS) servers, which can be purchased on a subscription basis, and can be integrated with existing IT infrastructure.


Nokia believes 5G operators will want to deploy a layered cloud architecture. This will include centralized and regional data centers, as well as high-processing capabilities deployed at the network edge – closer to where traffic is generated and where space is traditionally limited.


In April 2018, Nokia launched the AirFrame open edge server, which is designed to provide distributed computing capacity for far-edge data center deployments. The vendor says the open edge server will be fundamental to supporting MEC and ultimately 5G, where it can support ultra-low latency and time-critical applications such as industry automation, AR, and Cloud-RAN.


The open edge server incorporates hardware and cloud infrastructure software, including Nokia's Data Center Manager and cloud infrastructure for IoT and real-time applications, as well as professional services for the design stage. It can support enterprise applications or a private LTE network, as well as public network deployments. It is upgradeable with Nokia's own ReefShark chipset for optimized Cloud-RAN performance.

One of the first customers for the open edge server is MIRIS, a Norwegian real estate and technology firm, which will deploy the technology to support smart city services in business parks and residential areas. MIRIS plans to build data centers in about 20 urban locations in Norway during 2019, followed by a wider rollout across the Nordic region.


At MWC 2019, Nokia and China Mobile demonstrated the use of edge cloud combined with AI and ML to support real-time VR gaming. The PoC employed an open RAN architecture and was hosted on Nokia's real-time edge cloud computing platform, including the open edge server to provide the required real-time computing capabilities for VR video rendering in the cloud. This ensures ultra-low latency, no jitter, and adapts to the highly variable bit rate of the VR gaming traffic, says Nokia.


Ericsson

Ericsson is pursuing its own vision for edge computing based on the concept of a distributed cloud that unifies the centralized and decentralized resources in service provider networks. The vendor is working with operator partners and application providers to build an ecosystem based on Ericsson's network platform, the Unified Delivery Network (UDN), which has 22 core locations worldwide serving 85 service provider partners, and provides a global footprint for its edge cloud offering. In pursuit of its goal, Ericsson has set up Edge Gravity, a separate business within its Technology and Emerging Business unit. The aims, approach, and business model of Edge Gravity encompass the following:

  • building a web-scale cloud platform, the UDN Edge Cloud Platform, a one-stop shop for service, application, and content providers to quickly enable solution offerings on a global basis
  • using the cloud platform to enable experimentation with a variety of technologies to determine optimal techniques for future edge computing services
  • connecting service providers with applications running on the network and help ensure their role in the emerging global network for edge computing
  • exposing and monetizing the infrastructure back to the internet application ecosystem to establish a new way for the service provider community to participate in the internet services value chain
  • addressing two key segments – infrastructure-as-a-service (IaaS) and platform-as-a-service (PaaS) – representing a combined total addressable market of $90bn according to Ericsson
  • focusing on serving applications that require a distributed global reach, very low latency and high resilience, localized processing, real-time analytics, and data sovereignty.


Edge Gravity has a two-sided business model whereby, in return for allowing it to deploy its Cloud Platform deep inside of their networks, service providers receive a revenue share from revenue generated by provisioning customers' internet applications and from monetizing the use of the infrastructure, platform, data, and its related services.


Under an agreement with Limelight Networks, which concluded in October 2018, Limelight's content delivery technology will be deployed on the Ericsson Edge Cloud Platform to expand its global delivery capabilities and increase delivery capacities.


Ericsson believes that edge computing deployments will involve managing different types of sites, with the location of the edge being dependent on the use case and the functionality required to support it. For example, in an AR use case, the "edge" has to be located at different places in the network in order to meet both technical and financial requirements.


Ericsson is actively working with operators on edge computing trials, PoCs, and 5G deployments, including the AT&T Foundry Edge Computing Test Zone and KT's IT edge cloud. Though not a member of ETSI MEC, Ericsson participates in a number of industry groups and initiatives around edge computing, including the Automotive Edge Computing Consortium.


Ericsson is also a partner alongside NEC, Nokia, InterDigital, Orange, Telefonica, and academic institutions, in 5G-Transformer, a project established under the EU's Horizon 2020 research program and 5G Public Private Partnership (5G-PPP). The aim of the project is to design, develop, and trial solutions for CSPs and industry verticals based on NFV/SDN and MEC.


In February 2019, Ericsson announced a tie-up with Intel spanning open cloud and NFV infrastructure, from centralized data centers to the edge, and aimed at allowing telcos to take advantage of multivendor hardware options, Ericsson's end-to-end software solutions, and Intel's latest architectural solutions.


Huawei

Huawei launched its MEC@CloudEdge mobile edge computing platform in 2016, primarily targeting 5G networks. At the time, Huawei said that current mobile technology, including LTE, could not minimize latency even though the air interface throughput is increased tenfold, as end-to-end latency is improved only threefold due to the sub-optimal network architecture.


Huawei's MEC solution encompasses applications, content, core network service processing, and resource scheduling functions at the network edge, close to the RAN. According to the vendor, its major advantages are:

  • a cloud-based architecture that delivers enhanced network reliability and performance through the use of a cloud-native software architecture with stateless design and cross-domain resource sharing
  • a network architecture featuring control and user-plane separation (CUPS), allowing control-plane functions to be deployed centrally while user-plane functions can be flexibly deployed at the network edge to be close to users (Huawei has been deploying CUPS commercially since 2018)
  • intelligent application integration and service chain orchestration to provide rapid onboarding of new services and local processing of content and services
  • extreme forwarding capability to support big traffic local breakout and massive IoT connections
  • an open platform that supports compatible virtual machine and container ecosystems. This allows rich third-party application integration without coding.


The vendor says its MEC solution can be used in a wide variety of 5G scenarios, such as autonomous driving, AR maintenance, and telemedicine, and can be deployed flexibly in hotspot areas, such as enterprise campuses, stadiums, and central business districts.


The Huawei MECsolution allows edge nodes to provide intelligent services nearby and implements network management in the cloud. Huawei expects innovation around MEC to open up new markets, including B2C (mobile CDN), B2B (enterprise IT services and smart factory), and B2B2X (intelligent transport systems using AR and LTE-Vehicle – LTE-V – technology). It says that latency is not always the main issue for MEC, and that savings on backhauling costs and loading on the core network are also important. It believes MEC can enable applications such as IPTV over fixed wireless access networks to provide triple-play services, and is working on a number of MEC use cases, including a smart factory solution with China Telecom and field trialing a local data package for a smart museum application.


Huawei says it is dedicated to promoting MEC's open ecosystem development and actively supports MEC industry alliances. The company is closely engaged in the standards process for MEC, having been instrumental in establishing the Release 15 5G MEC activity within 3GPP, the introduction of MEC into vertical industries as part of Release 16, and the creation of an enhanced 5G architecture to support MEC. Huawei was also an ETSI MEC ISG startup member.


Samsung

Samsung participates in building 4G and 5G networks in major markets, including Korea, the US, Japan, and India. The vendor has already launched commercial 5G services with operators in the US and Korea, providing 5G equipment from radio frequency (RF) and modem chipsets to network core and radio solutions, and devices. Samsung is a member of ETSI MEC and is actively engaged in research into the application of edge computing in areas such as the connected car market, as well as pursuing its goal of building a strong presence in the IoT ecosystem for connected devices spanning the home and industrial applications.


Samsung is trialing MEC architectures designed to guarantee minimal latency and information security for autonomous driving and V2X use cases covering both 4G and 5G. In December 2018, the vendor signed an MoU with the Korea Transportation Safety Authority (KOTSA) to build a test bed for autonomous driving technology in Korea. The facility will also serve as an innovation lab for industry partners. Samsung believes MEC resources located near base stations are critical in supporting instant communication between fast-moving vehicles and will enable cars to react to unanticipated events.

Samsung says it is actively collaborating with and exploring the business opportunity for MEC technology with operators in Korea, as well as with other global operators.


The company's attempts to expand its smart-home IoT offering suffered a setback at the end of 2018 with the closure of its Artik products business. The Artik edge gateway modules featured high-performance processing power and provided a hardware platform targeting OEMs and manufacturing customers that need to process large volumes of data locally for latency, cost, or reliability reasons.


ZTE

ZTE MEC solutions address a range of service scenarios and use cases for 5G, including service localization, local caching, V2X, and IoT. As well as providing edge infrastructure, the vendor says it can help to accelerate edge application development for both operators and third parties wanting to build edge vertical applications.


Support for industry vertical use cases is critical for the large-scale deployment of edge computing, according to ZTE, and the vendor has focused its efforts on edge service implementation in these kinds of application scenarios from the outset. Founded on operators' wireless network and service capabilities, such as high-precision indoor positioning, intelligent network optimization, and video acceleration, the vendor addresses four major industry application fields:

  • big video (including live streaming and cloud gaming)
  • smart manufacturing (such as quasi-private networking, real-time control of field devices, remote maintenance and control, industrial HD image processing, and other industrial applications)
  • smart grid and smart city
  • V2X or "Internet of Vehicles."


ZTE believes operators will need edge cloud for 5G, but says the case for MEC is broader than just reducing latency, and that there are many applications that will require MEC for other reasons, including content caching and savings on cost and capacity of backhaul.


Due to the complex architecture of the edge computing platform, and the numerous technologies involved, ZTE supports the need for a standardized and unified, open edge computing architecture and platform for APIs. The vendor believes that the most effective business model for operators wanting to deploy edge computing will incorporate both public and private cloud, combining the technical and financial advantages of large-scale public cloud provider resources with a telco-enabled, highly efficient, lightweight edge cloud network, offering features such as ultra-low latency, enhanced throughput, and local caching to support applications such as live video streaming, AR/VR, and gaming.


ZTE offers a range of MEC servers for various application scenarios, from traditional COTS

rack/blade servers to units designed to be deployed at the base station. During MWC 2019, ZTE released its ES600S MEC server, which incorporates AI-based intelligence at the network edge to facilitate high-volume, real-time applications such as facial recognition and analysis of multiple video streams.


ZTE is working on a range of edge-computing pilot projects in China, including security, surveillance, and remote supervision for the port of Tianjin, video analytics, and surveillance for the campus of Zhengzhou University, as well as use cases supporting intelligent manufacturing, auto-parking, retail, smart campus, video caching for tourism, and VR gaming, at various locations throughout the country. Far from delaying the adoption of edge computing, ZTE believes operators in China will employ the technology from the outset in their 5G networks – both for consumer applications such as VR cloud gaming and 4K video delivery, and for industrial use cases.

Author

Julian Bright, Senior Analyst, Network Infrastructure and Software

julian.bright@ovum.com


Ovum Consulting

We hope that this analysis will help you make informed and imaginative business decisions. If you have further requirements, Ovum’s consulting team may be able to help you. For more information about Ovum’s consulting capabilities, please contact us directly at consulting@ovum.com.