Marvell Validates 51.2 Tbps AI Ethernet Switching with Keysight
Case Studies
Keysight Technologies achieved a new milestone in validating the accelerated infrastructure needed to power next-generation, AI-driven cloud data centers, using Keysight’s AresONE-M 800GE test platform and the Marvell Teralynx 10 Ethernet switch chip.
AI networks require high performance, unprecedented scale, and low-latency switching fabric to handle massive volumes of data and the real-time decision-making advanced AI workloads demand. Keysight collaborated with Marvell Technology to address the testing needs of network equipment manufacturers, silicon vendors, and data center operators building and managing high-speed 800GE and 1.6T networks.
Without rigorous testing, product re-designs, recalls, or damage to customer trust become real risks. Marvell needed a way to definitively measure and validate the low-latency capabilities of its Teralynx 10 Ethernet switch chip, which is designed to support up to 51.2 Tbps of throughput — far beyond what most Ethernet test setups are built to exercise at full line rate.
Proving that switch silicon can maintain connection and low latency across dozens of 800GE links simultaneously, while measuring loss, latency, and jitter at line rate, required a testbed capable of scaling well beyond typical lab configurations.
The milestone testbed paired Keysight’s AresONE-M 800GE, the industry’s most comprehensive multi-speed Ethernet performance test platform and traffic generator, with the Marvell Teralynx 10 Ethernet programmable switch device, which offers industry-leading low latency and supports up to 51.2 Tbps of throughput with comprehensive analytics.
Eight AresONE-M 8-port chassis were chained together, providing an industry-first 64 x 800GE link configuration to build a high-scale testbed running at 800GE line rate. The AresONE-M generated 51.2 Tbps of traffic and sent it through the Teralynx 10 switch to test its limits.
The AresONE-M 800GE test platform validated an 800GE interface speed based on 112G SerDes, which facilitates faster data transfer between devices and extended reach in data center interconnects or telecommunications networks that run data-intensive AI applications.