ONES 4.0: How Enterprises Gain Control and Observability in AI Networks
September 16, 2025
ONES & SONiC
TL;DR:
- ONES 4.0 introduces NVIDIA SpectrumX orchestration, automated RMA, and AI Infra Analytics.
- New UX delivers real-time health, streaming, and utilization insights.
- Enhanced rule engine improves monitoring, alerting, and resiliency.
- A milestone release future-proofing AI network fabrics with intelligence and automation.
Introduction
ONES 4.0 marks a major step forward in the Open Networking Enterprise Suite (ONES) journey. Designed for the AI networking era, it brings:
- SpectrumX orchestration for AI clusters.
- Automated RMA for hardware replacement.
- A redesigned UX with real-time observability.
- AI Infra Analytics for GPU health and performance.
This release embodies our vision of building \"Networks for AI and AI for Networks\" delivering control, visibility, and confidence at enterprise scale.
What Makes ONES 4.0 Different?
ONES 4.0 integrates AI-ready orchestration, end-to-end observability, and automation into a single platform, helping enterprises simplify operations and ensure resilient networks.
How Does ONES 4.0 Support NVIDIA SpectrumX?
Figure 1: Nvidia Spectrum-X Multi tenancy orchestration using ONES
- Automates deployment from Day-0 design to Day-2 scaling.
- Works across multi-tenant, greenfield, and brownfield environments.
- Integrates with NVIDIA Air for seamless GPU/CPU/network management.
- Ensures reliable, optimized AI performance with simplified operations.
Figure 2: GPU Allocation for Tenants using ONES
How Does ONES 4.0 Simplify RMA?
Figure 3: RMA New Device Onboarding and Configuration restoration using ONES
RMA is now automated, with backups, restores, and history tracking handled by ONES.
- Eliminates manual backup/restoration.
- Full visibility into each replacement step.
- Faster recovery from hardware failures without disruption.
What is New in the ONES 4.0 User Experience?
Figure 4: Aviz ONES Centralized Orchestrator and Management
- Device Health Status → real-time operational insights.
- Device Streaming Status → telemetry and data flow visibility.
- Regional Network View → consolidated perspective across geographies.
- Capacity Utilization → track resource use.
- Critical Alerts → prioritized for faster response.
How Does ONES 4.0 Improve GPU Analytics?
Figure 5: GPU Health Analytics, reporting, and Monitoring
- GPU type & model identification.
- Health status breakdown for proactive monitoring.
- Utilization metrics (GPU, memory, power).
- Thermal management: temperature & fan monitoring.
- ECC error reporting for reliability.
This enables teams to monitor GPUs at scale ensuring performance, stability, and early detection of issues.
What is New in the ONES 4.0 Rule Engine?
VRRP visibility for HA monitoring.
New rule sets:
- Configuration sanity checks.
- GPU health monitoring.
- NTP sync failure alerts.
- Broadcast/multicast storm detection.
Strengthens automation, proactive issue detection, and operational resiliency.
Conclusion:
ONES 4.0 is more than an upgrade it is a platform for the AI era:
- SpectrumX orchestration for AI networking.
- Automated RMA for faster recovery.
- Unified UX for end-to-end observability.
- AI analytics and rule-based automation for resilience.
Ready to explore ONES 4.0?
Book a demo with Aviz and see how we are transforming AI-ready network operations.
