Broadcom, the tech conglomerate, has enhanced one of its flagship networking chips, Trident 5-X12 (AVGO.O), by integrating artificial intelligence (AI) features. The updated chip aims to optimize data movement within data centers, improving efficiency and facilitating tasks such as network security.
Optimizing Performance: AI Integration in Trident 5-X12
By allocating a portion of the silicon in Broadcom’s Trident networking processors to AI, the chip exhibits enhanced performance, including reduced power consumption and increased network bandwidth. Addressing challenges in managing massive AI clusters, the Trident 5-X12 chip is designed to alleviate network traffic congestion.
Robin Grindley, an executive in Broadcom’s Core Switching Group, highlighted the chip’s capability to tackle specific networked computing tasks, such as AI, where software alone proves too slow. The neural network component of the chip analyzes all packets and traffic patterns, identifying nuances beyond the capacity of standard approaches.
Activating AI Functionality for Effective Traffic Routing
Grindley explained that the AI functionality can be activated after a customer builds an AI model based on their data center traffic. Data center operators can train the model to detect unwanted traffic types, such as denial-of-service attacks or network congestion. Once constructed, a version of the model is integrated into the chip to assist in routing traffic effectively.
The decision to incorporate AI features was made approximately two years ago when Broadcom introduced programmability into a prior chip version. The Trident 5-X12 utilizes 5-nanometer manufacturing technology and is currently available for shipping to qualified customers.
Read More (AI)