关于AI网络架构的文章

思科OCP anounce了800G 51.2T G200-based minipack3 switch。对比之前Tesla anounce的TTPoE。真的很好奇，谁是AI-networking的未来，以及思科是否走在正确的路上，以及S1背后的技术。

大致浏览了相关的文章，先mark住，回头再细读一下。

OCP Summit 2024: The open future of networking hardware for AI - Engineering at Meta

Meta’s open AI hardware vision - Engineering at Meta

https://www.opencompute.org/documents/meta-minipack3-ocp-specification-all-in-one-for-review-pdf

Bytedance DDC:

https://drive.google.com/file/d/17d6lUJpgj7KqekuTtJSYJM6OUQJw5Sai/view
 

Tesla's TTPoE:

https://github.com/teslamotors/ttpoe

 Discussion:

https://www.linkedin.com/feed/update/urn:li:activity:7256314954897600512/

https://www.linkedin.com/in/kevin-f-10221826/recent-activity/all/

Demystifying AI Networking Protocols（randy zhang）

https://medium.com/@randyhzhang/demystifying-ai-networking-protocols-528cd29a51f6

Design principles of internet:

https://cs.ucr.edu/~ztan/courses/CS204/f23/lec2_principles.pdf

Queueing theory:

https://en.wikipedia.org/wiki/Queueing_theory

smart edge, dump core:

https://web.mit.edu/saltzer/www/publications/endtoend/endtoend.pdf

C8000 cisco live:

https://xrdocs.io/8000/blogs/cisco-8000-cisco-live-sessions/

S1 forwarding and QoS:

https://www.ciscolive.com/c/dam/r/ciscolive/apjc/docs/2023/pdf/BRKSPG-2208.pdf