Scaling network infrastructure for AI faces a daunting challenge. Compared to today's servers connected at 100 Gbps, AI clusters will demand 30 Tbps - a 300X increase.
Nigel Alvares, VP of Global Marketing and Business Planning, Marvell, explains:
- The current cloud AI infrastructure is facing a bottleneck issue due to the need for greater bandwidth connectivity between growing capabilities of GPUs, TPUs, and AI accelerators.
- The solution to this challenge lies in the use of Optics. Optics are needed to connect these clusters together, within the clusters, and to connect the AI engines to memory, storage, and to one another.
- Marvell is addressing this opportunity with PAM4 technology spanning 800 gigabits per second, aiming to reach 1.6 terabist per second over the next year. They are also working with the ecosystem to scale beyond this.
Check out other Tech Updates on our YouTube Channel (subscribe today): https://www.youtube.com/@NextGenInfra and check out our latest reports at: https://nextgeninfra.io/
Have a tech update that you want to brief us on? Contact info@nextgeninfra.io!
https://ngi.fyi/tech-marvell-nigel
