본문 바로가기 주메뉴 바로가기 검색 바로가기
Edge Computing Is Coming. But What Is It?
상태바
Edge Computing Is Coming. But What Is It?
South Korean tech companies and mobile carriers are already pushing ahead with the new technology for faster and cost-effective data processing.
PHOTOGRAPH: Unsplash
PHOTOGRAPH: Unsplash

By Jenny Lee WIRED Korea

The year 2019 was the year some parts of the world got their first taste of 5G through commercial networks and 5G-enabled devices. But this year and the next will usher in a full-blown 5G era, delivering its initial promises of ultra-high speed, low latency and massive connectivity.

That was the prevailing view expressed by industry leaders at the world’s largest Consumer Electronics Show (CES) held early January, where they demonstrated exciting real-world 5G applications -- including artificial intelligence (AI) tools, smart devices, connected cars, drones, robots and more.

With a remarkable progress made on the connectivity agenda with 5G, technology’s influence on human behavior will move forward in this new decade, said Gary Shapiro, president and CEO of the Consumer Technology Association, which organizes CES.

One of the most anticipated benefit of 5G is the ability to seamlessly connect embedded sensors in virtually everything. Analyst firm Gartner projects that enterprise and automotive IoT endpoints, which include end devices like sensors and gateways, will climb to a total of 5.8 billion by 2020, a 21 percent increase from last year.

And the explosive growth of IoT devices will start generating unprecedented volumes and variety of data at an unprecedented speed and create new demands to rethink how data is ingested, stored, processed and analyzed.

While all data is currently handled in a centralized data center or a cloud, but edge computing -- formally known as multi-access edge computing (MEC) -- is suggested as a more efficient alternative today.

The basic idea of edge computing is processing large amounts of data near “the edge of a network,” closer to where data is being generated and sending only the data that needs further analysis to the cloud. Because data no longer has to traverse over a network to a central location for processing, network latency and congestion would be significantly mitigated, thereby allowing end devices to perform faster and better, said Funfun Yoo, head of XS Lab, a Seoul-based company that specializes in developing edge computing solutions.

Such filtering of data closer to the sensors can slash costs for internet bandwidth usage and cloud data storage. While other benefits abound, the technology can be particularly valuable for tasks that do not tolerate delays of any kind like self-driving cars and healthcare monitors.

“If the IoT use case requires sub-second response time for safety and critical operations, then waiting for a response from the cloud may not be viable,” said Yoo, who is also an adjunct professor at Hanyang University’s College of Computing. “The processing of anomaly detection by a sensor installed in, for example, a chemical plant and the processing of the decision to stop a machine must be carried out immediately as a delayed response time may cause serious human damage.”

“Processing the data at the edge is a way to achieve sub-second response times,” Yoo said.

Despite huge costs involved in setting up an edge computing infrastructure, there is not a lack of interest nor activity in edge computing, in which many vendors see potential business opportunity.

Small startups are producing specialized edge computing stacks, which are used to store a collection of data. IoT platform vendors have edge solutions connecting to IoT hardware, and other vendors are looking to launch a complete device to cloud solution.

Global tech companies including Google, Amazon, Dell, IBM, Intel, Microsoft and Oracle are taking a deep dive into edge computing. Also joining the fray are South Korea’s IT giants as well as the country’s top mobile carriers, whose competition has become intensified this year.

“Telecom companies’ role is to build the infrastructure while cloud service providers like Naver and Kakao will work to connect the edge computing infrastructure to their cloud,” the professor said.

Last month, SK Telecom announced the launch of the Global MEC Task force with Bridge Alliance member operators, including Singtel, Globe, Taiwan Mobile and PCCW Global, to make joint efforts for the development of 5G mobile edge computing (MEC) technologies and services.

It said it was also setting international MEC standards to build an interoperable MEC platform.

If SKT’s goal is to create a pan-Asian 5G MEC ecosystem, KT’s goal is much broader. KT has also formed an alliance in January with five major global telecommunication firms to boost its MEC capability. The members of the 5G Future Forum include America Movil of Mexico, Rogers Wireless of Canada, Telstra of Australia, Verizon Wireless of the United States and Vodafone of Britain.

The alliance will focus on developing global specifications and standards for 5G MEC interoperability, according to KT, which has already set up MEC communication centers in eight major cities in Korea last year.

Market Research Future (MRFR) study shows that by 2024, the global market size of edge computing is expected to reach $22.4 billion.

Nevertheless, actual deployment of this technology is hinged upon both its maturity and its specified definition by relevant standardization bodies, industry groups and open source projects.

“The technology is still in its fledgling stage,” Yoo said. “We have many of the necessary technologies to make it happen but it is still a year or two away from becoming fully matured.”

이 기사를 공유합니다
  • Jenny Lee Staff Reporter jlee@wired.kr

    Jenny Lee covers information and communications technology, finance, and industry for WIRED Korea. She has previously worked with the Voice of America and the Associated Press Television News.