How CNCF Projects United Two Countries' Engineers to Solve AI's Power-and-Land Problem
summary
NTT DOCOMO BUSINESS and SK Telecom stitch eight data centers across more than 2,000 km over an IOWN APN wide‑area L2 network and operate them as a single Kubernetes cluster using KubeVirt VMs so existing workloads run unchanged with tenant isolation. Measured inter‑GPU hits saturate the 100 gig wire rate and intersite RDMA storage reaches more than 10 GB/s. The distributed GPU fabric is built and proven in production.
AI-written summary · about
related talks
25mHow We Power the Largest AI Deployments on the Planet: Running Vir...
15mKeynote: From Platforms to AI Factories. Has Kubernetes Solved It?
05mSponsored Keynote: Scaling Global Innovation with AWS and the Kubernetes Community
10mApproach to AI at a large scale
09mAI for NI
16mAI usage in telcos
06mSponsored Keynote: AI-Ready Platforms: Scaling Teams Without Scaling Costs
24mBuilding an AI-Powered, Paved Road Platform with Cloud-Native OSS