AFFORDABLE DISAGGREGATED STORAGE FOR KUBERNETES AI/ ML/DL DISTRIBUTED WORKLOADS HIGHLIGHTS > Efficient storage disaggregation supports AI/ML/DL distributed workloads using: > Excelero NVMesh ® Technology > NVIDIA Mellanox BlueField ® Data Processing Unit (DPU) with NVMe SNAPTM > Software-defined storage virtual volumes exposed to bare metal K8s workers as local NVMe drives using NVMe SNAP means: > K8s CSI is not required > Clientless solution for ease of deployment > Single shared volume that is easily exposed to a distributed set of workers without any performance penalties > Simple RoCE configuration > Fastest time to ROI in the market EXECUTIVE SUMMARY Artificial intelligence (AI) implementations based on deep learning (DL) continue to gain traction thanks to their

Join for free to read