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 ONES Automates Spectrum-X Fabrics from Design to Deployment

Blog date iconSeptember 23, 2025Blog category iconAI Factory
How ONES Automates Spectrum-X Fabrics from Design to Deployment

Summary:

  • Spectrum-X provides high-performance Ethernet fabrics for AI workloads.
  • ONES Fabric Designer automates Day-0 design, underlay config, and server integration.
  • Features include intent-based design, IPCLOS automation, RoCEv2 QoS, and server NIC optimization.
  • Result: validated, production-ready configs in minutes, not weeks.

Introduction:

Building an AI-ready network fabric is one of the most complex tasks in infrastructure engineering. With NVIDIA Spectrum-X providing the hardware backbone, the challenge shifts to orchestration: How do you translate design intent into a production-ready fabric without weeks of manual work?

That's where the Aviz ONES Fabric Designer comes in. It transforms Day-0 design into a fast, error-free, intent-driven process, automating IPCLOS topology, QoS, and server node integration.

How Does ONES Simplify Spectrum-X Fabric Deployment?

ONES automates design and configuration with intent-driven blueprints, ensuring validated, production-ready Spectrum-X fabrics that are lossless, resilient, and AI-optimized.

How Does Intent-Based Design Work in ONES?

  • Define high-level inputs: Scale Unit size, GPU count, IP ranges.
Configuring the number of SUs using ONES
Figure 1: Configuring the number of SUs using ONES
  • Auto-generate full fabric design and configs.
Spectrum-X fabric auto-generation
Figure 2: Auto-generated fabric design and configuration
  • Validate topology before deployment.

Outcome: A logical blueprint that eliminates manual IP/BGP mapping.

How Does ONES Automate IPCLOS Underlay?

  • Spine/Leaf configs: Auto-calculated per GPU/server count.
  • BGP automation: ASN assignment + peering configs.
  • Interface setup: Ports, IPs, MTU consistent across fabric.
  • Server NIC configs: RoCEv2 tuning, PFC/ECN, hardware offloads, adaptive routing.
Spectrum-X orchestration using ONES
Figure 3: Spectrum-X orchestration using ONES

What once took days of CLI work now happens in minutes.

Spectrum-X fabric design using ONES
Figure 4: Spectrum-X fabric design using ONES

How Does ONES Ensure Lossless RoCEv2 Performance?

  • Auto-applies best-practice QoS profiles for RoCEv2.
  • Consistent DSCP trust and buffer settings across ports.
  • Eliminates packet drops that can derail AI training jobs.

Spectrum-X fabrics are InfiniBand-class, out of the box.

How Does ONES Integrate GPU Servers and SuperNICs?

  • Auto-configures switch ports (VLANs, MTU, QoS).
  • Validates connections via LLDP.
  • Extends orchestration to BlueField-3 SuperNICs for end-to-end tuning.

Ensures GPUs connect flawlessly to the fabric with zero manual intervention.

Conclusion:

With ONES Fabric Designer, Day-0 Spectrum-X deployment becomes:

  • Intent-driven and blueprint-based.
  • Automated for IPCLOS, QoS, and NIC configs.
  • Validated via NVIDIA AIR before hardware deployment.
  • Zero-touch provisioned for fast, reliable rollout.

ONES turns fabric design into a repeatable, intelligent process that accelerates AI infrastructure deployment.

Want to accelerate Spectrum-X fabric deployment?

Book a demo with Aviz and see how ONES transforms AI networking from Day-0 to production.

Learn More

Share:FacebookTwitterLinkedIn
Tags:aviz ones NVIDIAnvidia spectrum-xONES Fabric Designer Spectrum-XDay-0 Design AutomationIPCLOS AutomationRoCEv2 QoSIntent-Based Blueprint