Pdx®：人工智能训练/推理——公共云

Needs •Cloud-hosted AI Training / Inference•Secure private connectivity to the cloud•Performant access to Large Language Models (LLMs)i.e., RAG•Scalable footprint to support growth Challenges •Cost-effective training & inference for AI Models•Speed transfer of large data volumes•Integrating public cloud infrastructure•Managing end-to-end AI workflows Actions •Establish direct connections to public cloud•Connect to your private infrastructure•Deploy your private infrastructure in colocation•Operationalize Digital Infrastructure Hub Benefits •Cloud-adjacent colocation reduces latency•Performant throughput reduces latency•Security and compliance for data•Scalability of hosted GPUs for AI processing 3.Bring Public CloudData to Private Storage 4.Enable TrustedEnd-User Requests 7.SD-WAN – Software-Defined Wide Area Network8.VPN – Virtual Private Network9.HPC – High-Performance Compute10.CPE – Customer Provided Equipment11.PPE – Partner Provided Equipment12.GPU – Graphics Processing Unit13.HD or Std – High-Density or Standard Cabinet14.BI – Business Intelligence Analytics Tools