On March 21, 2018, the CCBN2018 Thematic Forum officially opened. According to Ronghe Net | DWRH.net, representatives attending this year’s conference primarily came from radio and television regulatory authorities across China, broadcasting stations, internet audio-visual enterprises, film and television production institutions, smart hardware companies, investment firms, data research organizations, and scientific research and academic institutions—totaling over 7,000 participants.
At the morning thematic session on March 21, Academician Liu Yunjie of the Chinese Academy of Engineering delivered a keynote report titled “Development Trends and Prospects of Future Networks.” Below is the full text of his speech—
Liu Yunjie: Good morning, esteemed experts. It is a great pleasure to exchange ideas with you today. Although the title of my presentation may sound familiar, today I will focus specifically on the needs of China’s radio and television (R&TV) industry—particularly in Part III, where I will offer strategic insights and recommendations regarding how future networks can be applied and promoted within the R&TV sector.
The Internet has now evolved for over 40 years, broadly progressing through three phases. The first 20 years were largely dedicated to academic research conducted at universities. Industry observers generally agree that the Internet evolves in roughly 20-year cycles—the first cycle being university-based research. The second 20-year phase (beginning around 1990) centered on consumer applications—such as e-commerce and WeChat—that dominate our daily lives today. With the emergence of the third-generation Internet—commonly termed the “Future Network”—we are no longer discussing some distant vision, but rather the network infrastructure we already have in place today. The name “Future Network” is admittedly somewhat misleading. Its defining feature is deep integration beyond the consumer domain—including industrial internet applications, as well as emerging ultra-high-definition video services like 4K TV and AR/VR. These new applications expose numerous limitations in today’s Internet architecture. Addressing these challenges constitutes the core theme of today’s discussion. The second-generation (consumer-focused) Internet was epitomized by the “3W” (World Wide Web) technology, which ignited explosive growth across all sectors—a dependency that persists to this day. So what are the core technologies underpinning the Future Network? There is broad consensus on this point, and I will elaborate shortly on what the Future Network is—and how we should understand it. A key technical characteristic of the next-generation network is software-defined networking (SDN), augmented by artificial intelligence (AI) and big data analytics—forming the foundational pillars of Future Network architecture. This represents our current understanding and observed evolutionary pattern. Similarly, mobile networks—also celebrating their 40th anniversary—follow a roughly ten-year cycle: 1G, 2G, 3G, and now the imminent transition from 4G to 5G, expected to launch commercial services around 2020.
The first major challenge facing the Future Internet is explosive traffic growth. As many know, Internet traffic growth has long outpaced Moore’s Law—and the current network architecture may soon reach its breaking point. This is Challenge #1. Challenge #2 stems from bandwidth- and latency-sensitive applications such as 4K/8K video and AR/VR. As noted earlier, industry forecasts suggest that 4K content alone could account for up to 80% of total traffic volume. Such applications impose stringent new requirements: high bandwidth and ultra-low latency. Huawei’s research report outlines specific performance targets—for example, 4K TV requires a minimum of 32 Mbps, rising to 75 Mbps for premium quality; 8K demands anywhere from 1 Gbps to 10 Gbps; and AR/VR requirements are even higher. Meanwhile, typical Internet latency currently exceeds 90 milliseconds. Clearly, these conditions present serious operational challenges. In response, Huawei has proposed “Network 5.0”—a dedicated initiative to address these demanding network requirements. Challenge #3 is the Industrial Internet—the convergence of the Internet with the real economy. This synergy underpins both China’s “Made in China 2025” strategy and global industrial transformation efforts. This is not merely a national priority; it is a universal imperative for all advanced economies. Integrating the Internet with the real economy is far more complex than consumer applications—but failure to clear this hurdle would severely compromise any nation’s long-term competitiveness. Consequently, countries worldwide are racing to seize leadership in this domain. U.S.-based GE and Germany’s Siemens are fiercely competing for dominance. For China, the stakes are exceptionally high: McKinsey estimated in 2016 that successful integration of the Industrial Internet with the real economy could yield $5.6 trillion in economic value. Why? Because unlike the “eyeball economy” of consumer internet—where value scales with user numbers—the real economy operates as a “value economy,” prioritizing productivity gains. China’s average labor productivity stands at only 15–30% of the OECD average—indicating enormous untapped potential. Thus, initiatives like “Internet Plus,” “Made in China 2025,” and the Industrial Internet are not buzzwords—they represent fundamental national strategies.
What kind of network infrastructure can meet these strategic requirements? Can ordinary Internet connectivity safely support factory-floor command, control, and scheduling? No one dares to try. Numerous accidents have occurred due to insufficient security, reliability, and service guarantees inherent in today’s public Internet.
So what exactly is the Future Network—and what is its objective? Does it require tearing down existing infrastructure and building an entirely new network from scratch? My view—and our practical implementation approach—is different. To illustrate: if today’s Internet were likened to a transportation system, it would resemble an ordinary road—even 100 Gbps or 10 Gbps links are akin to highways. By “ordinary road,” I mean that all users—whether leaders at Zhongnanhai or attendees here today—experience identical access speeds and security levels. Yet leaders require higher speed, greater security, and guaranteed priority—none of which the current Internet provides. In fact, even President Xi Jinping cannot be granted a dedicated lane on today’s Internet—unlike physical roads where VIP lanes exist. Why must we transform the Internet now? Because the emergence of the real economy, 4K/8K video, and AR/VR shifts the paradigm from human-centric to device-centric services—demanding ultra-low latency, ultra-high bandwidth, or dynamically shifting requirements (e.g., low latency for two hours followed by high bandwidth for the next two). Today’s static network simply cannot accommodate such variability. The solution lies not in replacing the “ordinary road,” but in overlaying intelligent, programmable “smart expressways” atop it—achieving flexible, on-demand resource allocation while preserving the underlying infrastructure. Without the ordinary road, there is no functional network—or transportation system. But this “expressway” is not conventional—it is intelligent, adaptive, and fully programmable. I believe this architecture directly addresses the needs of 4K/8K, AR/VR, and the Industrial Internet—and requires new architectures and new technologies to realize.
You may wonder whether this is feasible. In 2013, we deployed such a network in Nanjing—initially connecting four cities, later expanding to over a dozen. Built entirely on standardized hardware platforms, it leverages software-defined principles to flexibly provision diverse network slices: Industrial Internet, Energy Internet, and Vehicle-to-Everything (V2X) networks—all tailored to specific application requirements. Analogously, imagine dynamically creating dedicated lanes, expressways, high-speed rail corridors, or even air routes on an ordinary road via software instructions and virtualization. We have already achieved this. Over 80 domestic and international research teams are conducting experiments on this platform. CCTV’s “Xinwen Lianbo” (News Broadcast) has reported on it—so please do not dismiss this as speculative or futuristic technology. It is proven, operational, and commercially viable. Moreover, this work is not insular: we have published four papers in the prestigious ACM SIGCOMM conference—including two authored by our team—and three more in ACM CoNEXT. Our research is fully aligned with global standards. Developed nations—including the U.S. and EU—and international standardization bodies and industry consortia are all vying for leadership in this domain. Failure to secure this high ground means forfeiting global influence and voice.
I’ll briefly highlight several examples due to time constraints. New paradigms are actively shaping this field. Huawei’s “Network 5.0” initiative directly responds to the challenges outlined above. China’s three major telecom operators—China Mobile, China Unicom, and China Telecom—are also pursuing similar transformations. Yet, we see comparatively less strategic planning in China’s broadcasting sector. Today, I urge greater attention and action here. To build credibility, let me first cite international precedents before turning to domestic progress.
AT&T announced its “On-Demand Network Architecture” in 2014. By 2016, 30% of its traffic ran on the new architecture—exceeding the target at 34%. Its 2017 goal was 55%, though final figures remain pending. AT&T aims for 75% of all traffic to run on the new architecture by 2020. I’ll explain how shortly. Its CEO has declared that by 2020, AT&T will be predominantly a software company—not a traditional telecom operator. TBR Consulting surveyed 18 top-tier global telecom operators and found all had launched serious deployments. Their forecast projects a $158 billion market by 2021, growing at over 100% annually. This is a globally coordinated, high-priority initiative—not some distant future concept. How did AT&T implement SD-WAN in its backbone network? It recognized a fundamental shift: enterprises across all sectors—not just consumers—are migrating to cloud services. Thus, network architecture must pivot from carrier-centric to data-center-centric models. Public clouds, private clouds, and edge data centers are now distributed across major cities—including regional edge locations. How can a single network seamlessly interconnect them all? AT&T’s architecture solves precisely this: enabling unified delivery of public/private/edge cloud services, IoT applications, and enhanced security. This wide-area network now serves 150 countries—not just the U.S.
Regarding edge computing, AT&T repurposed its legacy central office facilities into CORD (Central Office Re-architected as a Datacenter) sites—transforming telecom closets into agile, cloud-ready infrastructure. Enterprises and end-users alike can deploy cloud services on demand. This market segment commands over $300 billion globally—nearly half of the entire telecom industry’s multi-trillion-dollar revenue base. The strategy also targets residential services: today’s proliferation of set-top boxes and home gateways will increasingly migrate to edge cloud platforms—not distant centralized clouds. Likewise, SMEs and large enterprises will offload compute, storage, and custom application workloads to either centralized or edge clouds—depending on latency, cost, and compliance needs.
Business orchestration forms another strategic pillar. Once deployed, network services can be automatically orchestrated and managed. AT&T is leading a joint initiative with China’s three major operators to standardize business orchestration frameworks. I believe China’s broadcasting industry possesses a rare opportunity to leapfrog ahead. While current broadcast networks lag behind telecom operators in maturity, they carry far less legacy baggage—making transformation faster and more agile.
AT&T is also integrating AI into its network operations. Given AI’s projected ubiquity within a decade, networks lacking intelligent automation will fail to keep pace with societal and industrial evolution—especially for drone fleets, autonomous vehicles, and robotics, all requiring intelligent, real-time network management. This is precisely AT&T’s ongoing work.
Now let me turn to China’s domestic efforts. Recognizing this global shift, China launched the “Future Network” Major Scientific Project in 2010—among the earliest national initiatives worldwide. Domestic research teams began exploring these concepts as early as 2005. In 2013, the State Council formally approved the Future Network as a National Major Scientific Project. Though initial enthusiasm led to intense competition among institutions vying for leadership, the project is now fully underway—with construction set to commence this year. The Jiangsu Provincial Government leads the initiative, with the Jiangsu Future Network Innovation Institute serving as the legal entity. Four core research units—Tsinghua University, the Chinese Academy of Sciences, the University of Science and Technology of China, and the Shenzhen Telecommunications Research Institute—are jointly executing the project. What scale does it cover? Approximately 40 cities nationwide and over 130 edge networks. Why such ambition? Because network sovereignty is integral to national strategy—directly impacting cybersecurity and functioning as the nervous system of the real economy. Sustaining technological leadership here ensures long-term competitive advantage and provides a national testbed for cutting-edge experimentation and pre-commercial validation—guiding industry-wide innovation.
Beyond these 40 cities, fiber-optic infrastructure is widely available. Many local governments—including those in southern Jiangsu, the Beijing-Tianjin-Hebei region, the Jiaodong Peninsula, and the Pearl River Delta—are proactively deploying next-generation network architectures. If executed successfully, China will occupy a globally leading position.
Let me outline several key platforms. First, we built an entire network using white-box switches—an architecture where commodity hardware replaces proprietary routers. Traditional routers cost hundreds of thousands of RMB per unit; white-box switches cost only tens of thousands—roughly one-tenth the price (though total cost includes transmission and other ancillary equipment). This hardware layer runs a network operating system (NOS) managed by centralized controllers. However, a single monolithic NOS cannot manage a national-scale network. Instead, we designed a hierarchical, multi-cloud architecture: separate NOS instances for backbone networks, data centers, and edge networks—coordinated under a unified control plane. Backbone bandwidth targets 100 Gbps, data centers 40 Gbps, and edge nodes 10 Gbps—enabling fully flexible, on-demand, user-driven resource allocation.
How can China’s broadcasting sector leverage this? This massive scientific project and its resulting innovations present an unparalleled opportunity—especially given our SDN/NFV-based, data-center-centric architecture. Broadcasting’s unique strength lies in its vast video content repositories—valuable, distributed data centers. Each province hosts dozens of large-scale data centers. How can these resources be shared nationally? SDN enables us to interconnect them into a unified national cloud—fundamentally distinct from telecom operators’ architectures and enabling entirely new service models. Your most valuable asset—your programming—is stored in these centralized or edge data centers. Interconnecting them via SDN—a mature, widely deployed technology—creates immense strategic value. A critical question arises: Can we develop our own network operating systems? Do we possess this capability? Yes—we’ve already developed NOS solutions for backbone, data center, and edge networks. Recently, China Unicom’s “A-Network” went live using this technology.
Previously, Unicom’s A-Network relied on heterogeneous hardware—part Huawei routers, part Cisco routers—each vendor’s equipment operating in isolation without centralized control. Adding a unified controller (essentially a software-defined brain) transformed operations: provisioning time dropped from one month to one minute. Once deployed, Alibaba and Tencent immediately requested access. This technology is now operational across over 300 Chinese cities—and has been running stably for nearly a year.
Similarly, QingCloud—a major Chinese cloud provider with four national data centers—replaced legacy routers with SR (Segment Routing) technology (an SDN variant) to achieve the same breakthrough. These solutions are fully domestically developed, secure, and controllable. Unlike mobile and PC operating systems—where China relies heavily on foreign vendors—network operating systems represent a new frontier where China can achieve full autonomy. Our controllers support heterogeneous hardware from multiple vendors—providing unified, vendor-agnostic management. Edge computing and central-office cloud solutions are likewise operational—deployed with telecom carriers and internet companies. Large-scale data center solutions are already powering Alibaba and Tencent’s infrastructure.
We’re also developing and deploying SDN-integrated content distribution platforms tailored for 4K/8K broadcasting requirements. Next steps involve leveraging AI and big data to further automate network maintenance and optimization—work already underway.
Finally, let me describe a demonstration project—implemented in two phases. Phase One focuses on 4K/8K and AR/VR applications, to be piloted in Guangdong Province. Provincial Governor Ma Xingrui is personally overseeing this initiative, collaborating closely with China’s broadcasting authorities—where they handle video encoding/decoding and we provide the underlying network infrastructure. Following Guangdong, the model will expand nationwide. AT&T’s network transformation was driven by similar strategic acquisitions and imperatives. If China fails to coordinate such integrated efforts, we risk falling behind again. 4K/8K, AR/VR, and the Industrial Internet represent a historic opportunity—one backed by a major national science project and strong local government investment. I urge all stakeholders—government agencies, equipment vendors, universities, and research institutes—to collectively seize this moment and elevate China’s broadcasting industry to world-class technological and service excellence. Thank you.
Last updated: 2026-03-08