HAS 2024 | Huawei Proposes the Industry's First L4 AI Security Intelligent Twin, Expediting Intelligent Defense for Network Security

[Shenzhen, China, April 17, 2024] During the 2024 Huawei Analyst Summit, Huawei proposed the industry's first L4 AI security intelligent twin at the forum themed "Accelerating Network Transformation Towards All Intelligence". This intelligent twin architecture offers new ideas and directions for global network security protection in the AI era, ushering in a new era of intelligent defense for network security.

Ma Ye, President of Security Product Domain, Huawei Data Communication Product Line

Ma Ye, President of Security Product Domain, Huawei Data Communication Product Line, delivered a keynote speech entitled "Industry's First L4 AI Security Intelligent Twin, Expediting Intelligent Defense". He expressed that with the accelerated advent of the AI era, the network security domain faces three challenges. First, with the proliferation in malware variants, it is difficult for generative AI technologies to effectively defend against malware. Second, it is difficult to detect encrypted attacks. Currently, up to 86% of attack traffic is encrypted. Such traffic has a high degree of covertness, and cannot be directly detected with traditional protection devices. Finally, it is difficult to handle massive amounts of events. Frequent attacks lead to a sharp increase in the number of security alarms. Manually handling these alarms is time-consuming, and so threats cannot be fixed in a timely manner.

To cope with the preceding challenges, Huawei proposed the industry's first network security autonomous driving model. Ma Ye believes that AI-powered security development will undergo five phases:

• L1: assisted defense, mainly conquering known single-point threats

• L2: low-level autonomous defense, automatically defending against new single-point attacks

• L3: medium-level autonomous defense, automatically defending against targeted attacks

• L4: high-level autonomous defense, automatically defending against new types of attacks

• L5: fully autonomous defense, ultimately achieving all-round autonomous defense against AI machine attacks

Currently, most players in the industry are in the L3 phase. Huawei has already entered the L4 phase, one-generation ahead of the industry. New technologies such as large models and graph AI are used, enabling autonomous defense against new types of targeted attacks and network-wide automatic security operations.

Huawei L4 AI security intelligent twin builds a three-layer defense architecture consisting of intelligent analysis, intelligent detection, and intelligent handling, fully upgrading intelligent network security defense. In terms of intelligent analysis, the graph AI intelligent security algorithm developed by Huawei can effectively detect 99.9% of variants and unknown malware. In terms of intelligent detection, Huawei's unique Encrypted Communication Analytics (ECA) algorithm and large model self-learning engine can detect 100% of malicious encrypted traffic within 30 minutes in decryption-free mode, five times faster than the industry average. In terms of intelligent handling, the security large model and massive high-quality corpus are used to implement automatic analysis, investigation, and handling of events without manual intervention, achieving 100 times higher operational efficiency.

Looking ahead, network attacks will remain rampant in the AI era. Huawei will stay committed to the R&D and innovation of fundamental network security technologies, continuously improve intelligent defense capabilities, and provide more reliable, efficient, and intelligent network security solutions for global users, thereby safeguarding intelligent transformation across industries.

For more information about Huawei's intelligent network security solutions, please visit https://e.huawei.com/en/solutions/enterprise-network/security.