| Tagline | 12.8 Tbps · 900ns Latency · 32× 400G in 1RU — The Enterprise & Cloud Data Centre Leaf/Spine Standard | 51.2 Tbps · 700ns · 64× 800G — Purpose-Built for AI/ML GPU Clusters and Next-Gen Data Centres | 21.6 Tbps · 24GB Ultra-Deep Buffer · 54× 400G — The Universal Leaf for SP, Cloud, and DC |
| Key Features | Up to 32× 400G (OSFP or QSFP-DD) or 128× 100G / 128× 25G in 1RU · 12.8 Tbps switching throughput · 5.3 Bpps forwarding rate · Sub-microsecond latency from 900ns — wire-speed L2 and L3 · 132MB intelligent integrated buffer with dynamic allocation | Up to 64× 800G OSFP800 (2RU) or 32× 800G (1RU) · 51.2 Tbps throughput · 21.2 Bpps forwarding (2RU model) · From 700ns latency — deterministic, line-rate performance · 165MB buffer (2RU) · 84MB buffer (1RU) — AI burst absorption | Up to 54× 400G ports (7280DR3A) or 72× 100G (7280CR3A) in 2RU · 21.6 Tbps throughput · 8.1 Bpps forwarding rate · Up to 24GB ultra-deep packet buffer with Virtual Output Queues per port · 400G ZR and ZR+ coherent DWDM optics — native DCI capability |
