Development Trends and Prospects of Future Networks

[[375296]]

Recently, Liu Yunjie, academician of the Chinese Academy of Engineering, dean of the Jiangsu Future Network Innovation Institute, and director of the Purple Mountain Laboratory of Network Communication and Security, published an outlook on "Development Trends and Prospects of Future Networks."

As the Internet moves from the consumer sector to the real economy, it is facing huge challenges. The traditional architecture of the Internet cannot provide deterministic service capabilities with low latency and low jitter. In addition, the development of 5G/B5G has also put forward new requirements for future network architecture: low latency, which urgently requires deterministic service capabilities; large bandwidth, which requires large bandwidth services that are sliced ​​on demand; and massive connections, which urgently require intelligent networking and management. The production-oriented Internet also puts forward differentiated requirements for the network. For example, in holographic communications, including ultra-high-definition video, AR, VR, etc., low latency, low jitter and other indicators are required.

Based on the challenges mentioned above, Academician Liu proposed a vision for the future network: In the future, any industry, any enterprise, and even individuals can customize the network according to their own needs. Operators will transform from traditional bandwidth sales to providing customized network service capabilities. In the future, the utilization rate of the entire network will increase from the current 50% to more than 90%, reducing user costs.

How to realize this vision? Academician Liu proposed a network architecture for 2030 - Service Customized Network (SCN). In it, the network operating system will control the operation and resource call of the entire network, and the protocols and addressing can be defined by the users themselves.

To achieve this goal, what technical problems need to be solved? First, the original closed network needs to be turned into an open network. The original resources are black boxes, and they need to be turned into open, transparent, definable, and reconfigurable networks. In other words, the data plane and the control plane need to be separated. The data plane needs to be reconfigurable, definable, and fine-grained. The control plane needs to be intelligently managed. On the business plane, cloud-network hyper-convergence and full cloudification of business are required.

Key Technology 1: Network Operating System

The network operating system is the brain of the future network and is becoming the commanding height of a new round of competition. We urgently need to seize this major opportunity for change and build a full-scenario, fully compatible, high-performance, and intelligent network operating system. The large-scale network operating system (CNOS) has these characteristics. It has three major functions: real-time status perception, intelligent resource scheduling, and on-demand service. CNOS has been running stably for more than three years in 400+ cities and 1,100+ nodes.

Key Technology 2: Programmable Network (White Box Device Operating System)

The white-boxization of backbone and edge networks has become an important evolutionary trend of future operator networks, and a complete white-box network ecosystem covering chips, ODM, hardware, and software has been initially formed.

Key Technology 3: Deterministic Network

Deterministic networks pose new challenges to traditional networks. From "best effort" to "punctual and accurate", flexible on-demand slicing, controllable forwarding paths, and bounded latency/jitter are the development visions of deterministic networks. Slicing, bandwidth, routing and other control technologies have been basically realized, and latency and jitter control are becoming the core technical issues in current deterministic network research.

Currently, the world is increasing its research on deterministic networks. Zijinshan Laboratory and Huawei have successfully conducted DIP experiments, using future network architectures and network operating systems, and have achieved major technological breakthroughs, releasing the results of the "world's first deterministic wide area network innovation experiment". In the Beijing-Nanjing thousand-kilometer test environment, the deterministic forwarding capability of DIP on the large network was verified, and the jitter can be controlled below 100 us, which can provide backbone network deterministic forwarding guarantee for industrial remote production control, remote surgery and other services.

Key Technology 4: Cloud-Network Integration

Cloud-network integration is also an important development trend and a top priority. The proportion of enterprises in China going to the cloud is relatively low, with only about 40% going to the cloud, while the United States has more than 80% and Europe has more than 70%. The reason for this phenomenon is that our cloud environment is not friendly enough. Now many users who go to the cloud need to go to multiple clouds, and users hope that after going to a cloud, all cloud resources can be shared. In this regard, we can already share cloud resources such as Alibaba Cloud, Tencent Cloud, and Huawei Cloud.

Finally, Academician Liu said that the future network is developing towards openness, white box, cloud, and SRv6, and the end-to-end capability and programmability will be further optimized and improved. The future network innovation test environment CENI will provide strong support for scientific research and industrial research in the network field, and can become one of the environments for operators' technological innovation and traffic exchange.

The first half of the Internet was dominated by Europe and the United States. Now, the Internet has entered the second half. Academician Liu hopes that we can use more than ten years of accumulation and precipitation to strive to make China's Internet go faster and better in the second half, and make China's contribution to the world Internet!