Cisco
Why DDN + Cisco
Turn AI infrastructure into AI advantage
DDN and Cisco bring together high-performance networking and data intelligence to create a fully integrated AI stack. This coordinated approach keeps GPUs fully utilized, removes data bottlenecks, and accelerates performance at scale—enabling faster time-to-insight and more efficient AI operations.
-
Accelerate Time-to-ResultsKeep GPUs fully utilized with a coordinated data and networking stack that eliminates stalls—reducing training times and speeding AI insights.
-
Eliminate Bottlenecks Across the AI PipelineCombine low-latency, lossless networking with high-throughput data pipelines to ensure seamless data flow and consistent performance at scale.
-
Deploy and Scale AI with Speed and ConfidenceLeverage pre-validated, modular architectures to accelerate deployment and scale predictably from pilot to production—without added complexity.
-
Increase Efficiency and ROI Across AI InfrastructureImprove job completion times, reduce operational overhead, and optimize resource usage to maximize the return on GPU and AI investments.
Cisco Nexus One for AI networking is fully validated with the DDN Data Intelligence Platform, which delivers high-performance AI data services through EXAScaler and Infinia. The combination pairs low-latency, lossless Ethernet with DDN’s data foundation so networking and data operate as one coordinated stack, keeping GPUs continuously fed at any cluster size.
Most AI clusters treat the fabric and the data layer as independent purchases, which is where hidden bottlenecks form. This is a jointly validated architecture: DDN delivers sub-millisecond latency, multi-terabyte-per-second throughput, and millions of IOPS, while Cisco contributes sub-microsecond switching latency and peak RDMA performance — together driving GPU utilization as high as 99%.
Yes. DDN provides native multi-tenancy, per-tenant isolation, QoS and SLA enforcement, and strong data protection, while Cisco adds hardware-based VXLAN Tunnel End Point isolation to keep client, workload, and storage traffic in separate security domains across shared infrastructure. This suits shared AI factories, regulated industries, and sovereign AI programs.
Can the DDN Data Intelligence Platform integrate with existing Cisco infrastructure?
It does. DDN supports POSIX, NFS, SMB, and S3 access across on-premises, hybrid, and multi-cloud deployments, and Cisco Nexus Dashboard manages storage, frontend, management, and inter-GPU backend networks from a single point — with RoCEv2 visibility, congestion analysis, AI job visibility, and anomaly detection.
Yes. Cisco and DDN have published a validated design for running Infinia as a software-defined storage layer on Cisco UCS M8 rack servers, with Nexus networking and Cisco Intersight management. This decouples the data intelligence layer from proprietary storage hardware, eliminating I/O wait and maximizing GPU utilization on Cisco compute.
Enterprises, neoclouds, cloud service providers, and sovereign AI programs building or scaling GPU clusters. The architecture is compliant with Cisco Enterprise Reference Architecture for clusters up to 1,024 GPUs, Cisco Cloud Reference Architecture for larger deployments, and the NVIDIA Cloud Partner reference architecture at any size.
Contact an AI Infrastructure Expert
We’re here to help you with all your AI data needs. Reach out to our team of experts for personalized support, inquiries, and solutions tailored to your specific requirements.
- Expertise and Innovation
- Tailored Solutions
- Dedicated Support