Abstract
Existing congestion management techniques for artificial intelligence (AI) data centers typically involve dedicated hardware for managing network congestion, which can be costly to implement, thereby requiring a large investment by network operators. Proposed herein is a technique to facilitate packet replication, which can make communications in an AI fabric resilient against congestion.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Ali, Zafar; Camarillo, Pablo; Filsfils, Clarence; and Abdelsalam, Ahmed, "PACKET REPLICATION FOR AI TRAINING IN EXISTING DATA CENTERS", Technical Disclosure Commons, (February 03, 2026)
https://www.tdcommons.org/dpubs_series/9267
COinS