[Shanghai, China, September 19, 2025] During the “AI-Powered WAN, Digital-Intelligent Future” WAN roundtable held at HUAWEI CONNECT 2025, Huawei unveiled its newly upgraded Xinghe Intelligent WAN Solution. This solution empowers customers to accelerate network monetization and industry intelligent transformation by building an IP network foundation with deterministic experience.
John Cai, President of Router Domain, Huawei Data Communication Product Line
John Cai, President of Router Domain, Huawei Data Communication Product Line, outlined four critical challenges facing Wide Area Networks (WANs) today, driven by surging video traffic, deepening digitalization, and the rise of quantum computing and intelligent computing services. First, as more than 170 countries push ahead with their digitization strategies, network convergence and the elimination of data silos are now imperative. Second, emerging services like cloud gaming and telemedicine pose higher requirements on differentiated bandwidth and latency. Third, with inclusive AI, a packet loss rate of just 0.1% can diminish computing efficiency on traditional networks by 50%, necessitating urgent improvements in transmission efficiency. And fourth, breakthroughs in quantum computing technologies pose severe threats to traditional network security.
To address these challenges, Huawei’s Xinghe Intelligent WAN Solution builds a deterministic network from four dimensions—connectivity, experience, computing power, and security—and features four core capabilities:
All-service convergence: In industry data network scenarios, Huawei leverages slicing and SRv6 technologies to ensure service isolation among government departments and reduce provisioning time to minutes. In industry production network scenarios, the built-in pulse code modulation (PCM) technology supports multi-rate interfaces with speeds ranging from 64 Kbps to 400 Gbps, facilitating smooth evolution of traditional SDH networks to IP ones.
Deterministic network: In industry data network scenarios, AI-powered application identification technology achieves over 95% accuracy in identifying encrypted flows, enabling automatic detection of video conference freezes and acceleration of mission-critical services. In industry production network scenarios, the AI-based latency compensation algorithm controls jitter at the microsecond level, and high-precision clock and multi-fed, selective receiving technologies ensure zero service misoperations.
High-computing-efficiency WAN: Huawei’s Xinghe series routers incorporate a flow awareness engine that identifies 99% of elephant flows. And global intelligent scheduling boosts network throughput from 40% to 90%. The Starnet lossless algorithm achieves zero packet loss during transmission over distances up to 1,000 km, reducing computing efficiency loss to below 5%. What’s more, the Starnet vector engine upgrades data transmission modes to safeguard sensitive information.
Intrinsic security: Quantum-resistant encryption and Xsec enable simplified deployment, creating a robust barrier against potential quantum threats. This meets the confidential transmission requirements of industries such as government and finance.
At the roundtable, several customers and partners spoke about the achievements of their collaborative practices. For example, Mohammad Mustaghfirin, CTO of Surge Indonesia, explained how the IPv6 Enhanced network facilitates Internet services along Indonesian railways. Listing Surge as the world’s first benchmark for self-built broadband operations in transportation, he revealed the company’s plan to build an IPv6 Enhanced target network that will deliver high-quality connectivity. Kenneth Bagarukayo, Commissioner of Data Networks at Uganda’s Ministry of ICT & NG, highlighted how Huawei’s SRv6 capabilities on NetEngine routers have reduced deployment time per MDA from 2 hours to just 1 minute. He also noted that network slicing allows critical services such as customs, tax, and banking operations to run over a single network. And Zhang Xianguo, General Manager of Network R&D Department at DYXnet, revealed that through collaboration with Huawei, they jointly developed an AI-native ultra-connectivity architecture. This architecture enables tenants to achieve elastic scaling from Mbps to 100 Gbps and significantly accelerates long-distance file transfers using RDMA, with speeds over 20 times faster than traditional TCP.
Looking ahead, Huawei will continue to deepen the integration of AI and networks, driving more industries toward intelligent transformation with the Xinghe Intelligent WAN Solution. The company aims to build intelligent, efficient, and reliable deterministic network foundations for global customers, reinforcing the network backbone for the intelligent era.