Aviz Logo
Contact
Modern network stack built for the AI era.
Modernize your Network
Start here: Cisco, Arista, Juniper network users; Gigamon and Netscout observability users
Autonomous Agentic AI PlatformNetwork CopilotAviz Packet BrokerAviz Service NodeVirtual Aviz Service Node / vTAP
SONiC & AI Experience Hub
AI for Networks. Networks for AI.
Hamburger
Aviz Logo

How Aviz ONES and NVIDIA Spectrum-X Automate Multi-Tenant AI Networking at Scale

Blog date iconSeptember 22, 2025Blog category iconONES & SONiC
How Aviz ONES and NVIDIA Spectrum-X Automate Multi-Tenant AI Networking at Scale

Summary:

  • AI workloads demand high-performance fabrics like NVIDIA Spectrum-X.
  • Aviz ONES automates design, deployment, and Day-2 operations with intent-based control.
  • Features: digital twin simulation, tenant segmentation, GPU-aware QoS, and agentless observability.
  • Together, ONES + Spectrum-X deliver reliable, lossless, and scalable AI networking.

Introduction

AI workloads are pushing data centers to new levels of performance. NVIDIA Spectrum-X provides a high-performance Ethernet fabric built for GPU-intensive environments. But building and managing such fabrics requires more than hardware.

That's where Aviz ONES (Open Networking Enterprise Suite) comes in, providing automation, observability, and AI-assisted operations that transform the entire Spectrum-X lifecycle, from design to Day-2 operations.

How Do ONES and Spectrum-X Simplify AI Networking?

ONES Orchestrating Spectrum-X Fabric

Figure 1: Orchestrating the Spectrum-X fabric using ONES

ONES automates Spectrum-X fabrics with intent-based design, multi-tenant management, and end-to-end observability, delivering predictable performance at scale.

How Does ONES Simplify Spectrum-X Fabric Design?

  • Blueprint-driven design → Define inputs, auto-generate configs.
  • Pre-deployment simulation → NVIDIA AIR digital twin validates before rollout.
  • Lifecycle control → Config mgmt, drift detection, rollback, version comparison.

How Does ONES Support Multi-Tenant GPU Fabrics?

ONES Tenant Segmentation

Figure 2: Tenant Segmentation using ONES

  • Secure segmentation with EVPN, VXLAN, VRFs.
ONES Tenant GPU Allocation

Figure 3: Tenant GPU Allocation using ONES

  • GPU-aware provisioning aligning network + compute.
  • Policy-driven QoS/PFC for lossless, congestion-free RoCE traffic.

How Does ONES Deliver Observability Without Agents?

  • RoCE/RDMA metrics → PFC, ECN, queue utilization.
  • Unified dashboards → Works across vendors/OS.
  • Proactive alerts → Integrated with Slack, ServiceNow, Zendesk.

No extra agents required, simplifying operations without losing depth.

What's the Advantage of Spectrum-X + ONES?

Capability Value Delivered
Fabric Automation Blueprint design + digital twin simulation
Multi-Tenant Control Secure EVPN/VRF isolation, GPU-aware QoS, PFC
Observability Agentless telemetry, unified dashboards, proactive alerts

Operators get faster deployment, higher reliability, and adaptability to evolving AI workloads.

Conclusion:

Aviz ONES + NVIDIA Spectrum-X = AI-grade networking that is:

  • Automated from design to Day-2.
  • Multi-tenant and GPU-aware.
  • Observable and reliable without agents.
  • Scalable for evolving AI demands.

Together, they make AI networking predictable, lossless, and easy to operate at scale.

Want to see how ONES + Spectrum-X can simplify AI networking?

Book a demo with Aviz and explore the AI-ready fabric advantage.

Learn More About ONES for NVIDIA Spectrum-X

Share:FacebookTwitterLinkedIn
Tags:aviz ones NVIDIAnvidia spectrum-xONES Spectrum-X AutomationDigital Twin SimulationTenant SegmentationGPU-Aware QoSIntent-Based Design