Multi-Access Edge Computing (MEC)

Definition

Multi-access Edge Computing (MEC) is a typical edge computing technology. It provides cloud computing capabilities and an IT service environment at the edge of a network, enabling real-time service response at the edge. As ICT convergence deepens, new services and applications, such as IoT, AI, and digital manufacturing are demanding better network capabilities. They require networks to provide low latency, strong compute power, and more secure connectivity. MEC is the key technology to address these challenges.

MEC runs on a combined hardware and software platform. It provides an IT service environment and cloud computing capabilities at the edge of mobile networks to reduce latency for network operations and service interaction. The MEC is close to the network edge of "things" and "data sources", and is an open platform integrating core capabilities, such as network, computing, storage, and application. The MEC provides nearby edge intelligent services to meet key requirements of industry digitalization in terms of agile connection, real-time services, data optimization, intelligent applications, security, and privacy protection.

Background

• Rise of IoT: Rapid development of IoT will place a massive number of its end devices into a lot of industry sectors, including industrial manufacturing, agriculture, education & health care, transportation & energy, finance & information, and home & environment. These IoT end devices require local service access and will produce massive amounts of traffic, jamming the networks and compromising user experience. Sure, carriers can add more resources in their data centers to handle the traffic increase. But considering the speed and scale of IoT development, telecom networks still need a more sustainable solution. Smart devices are at the edge of a network. They are various in type and usually have the compute and storage capabilities required to handle some services locally. If IoT services are deployed at the network edge, they can be processed and terminated locally. This greatly reduces the impact of massive terminals on transport networks and data centers, and maximally fulfils service requirements.
• Development of AI Technologies: The advancement of AI boosts the development of new use cases, such as self-driving, remote surgery, robot assistance, AR/VR, and live streaming. Some of them are particularly sensitive to latency requiring less than 10ms. Take self-driving and remote surgery as examples. They would require a latency within 1ms, which is impossible on conventional networks, as information needs to be transferred from the user devices all the way back to the core for processing, and then back to the user devices. If these new use cases are hosted by a network with a conventional architecture, the latency would be unacceptable and could even be catastrophic. MEC is one of the key technologies that 5G networks use to produce ultra low latency. While the central data centers take care of the analysis and processing of long-term, latency-insensitive data, MEC on the edge can handle data processing in proximity to end users, providing faster response and better experience.

• Digital Manufacturing: Modern manufacturing relies greatly on its digital base and big data analysis. Conventional isolated information silos are no longer sufficient for maintaining production efficiency and quality. Enterprises are in need of a system that can aggregate, merge, analyze, and process their data to facilitate automated production. Besides providing powerful data aggregation and processing capabilities, the system must also be highly reliable and capable of securing integrity, privacy, and effectiveness of production data. To be specific, the system must meet the following requirements:

• Keeping data local: Critical and confidential data must be analyzed and processed locally within the campus.
• Visual manufacturing: High definition visual data is collected to facilitate automatic measuring, positioning, and inspection. Monitored process statistics are quickly analyzed and returned to the control. Visual data is collected via industrial cameras and, usually, a fixed line network. A facility owner must redo the cabling and networking every time there is an adjustment or a new production line, which is time consuming and expensive. Using a wireless network will address the issue for good.
• Data filtering: A massive amount of data is generated in the manufacturing process, including, for example, the video content analysis system. Without proper filtering, unnecessary data will flood the network and overload the data processing unit. Raw data must be filtered at the edge before being transferred to the network. MEC is capable of fulfilling all requirements by providing storage, compute, and service processing capabilities from base stations or directly from factories. In this way, data can be processed locally and kept out of outside networks.

• Limits of Network Construction: Mines, docks, and ports require video surveillance to facilitate remote machine operation, protection over human, environment, and assets, and statistics-based production management. However, it is difficult and expensive to deploy and maintain a fixed line communication network at these locations. Wi-Fi is an option but risks service interruption, because it is not powerful enough to penetrate through obstacles, especially when made of metal, which is common. Mobile network base stations provide a better solution for locations without a fixed-line network. A combination of base stations and the MEC enables a mobile network to provide ultra-low latency and high-bandwidth capabilities to customers, fulfilling their needs on industrial IoT and stable remote machine operation.
• Service Development Trend: MEC applications are still at the kick-off stage. But with the fast development of the 5G networks and relevant fields, there is no doubt that MEC will be a strong anchor for telecom operators who plan to explore the B2B and B2C markets, and are ready to transform from connection providers to compute service providers.

• Cellular network replacing Wi-Fi: Large- and medium-enterprises usually build their own wireless networks and rely mostly on Wi-Fi. However, Wi-Fi is vulnerable to attacks and interference, and lacks stability and reliability, all of which can compromise production. 5G is a good option for enterprises who suffer from these issues. It provides secure, reliable, and stable services, as well as the MEC capabilities, helping enterprises build a fast-responding system and keep their data processing local.
• Boosting HD Video and VR development: By its nature, 5G provides higher bandwidth and lower latency, enabling the development of HD video and VR. With MEC, the content delivery networks (CDNs) are being moved to the edge, improving experience and boosting the popularity of HD video and VR services.
• Exploring high value B2B and B2C markets: Network slicing and MEC enable telecom operators to provide differentiated and tailored service experience through vertical industries. With the fast development of end devices, telecom operators are getting ready to build a fully connected ecosystem and explore more business opportunities.

Industry Applications

• Seven MEC application scenarios defined by ETSIThe ETSI has released the MEC service scenarios, including intelligent video acceleration, augmented reality, enterprise traffic steering, connected vehicles, Internet of Things (IoT), video stream analysis, and assistance for intensive computation. These scenarios will drive three trends: content processing at a regional level, app traffic steering with local breakout (LBO), and computing nodes moved to the network edge.

• Detailed MEC application scenariosThe following figure shows the detailed MEC application scenarios for individual users, enterprise users, IoT services, and other vertical industries. MEC satisfies experience requirements and also brings more business innovations for customers.