Aviz Networks ONES 4.2 details Day-0 automation and monitoring updates

Aviz Networks’ ONES 4.2 release expands the Open Networking Enterprise Suite with Day-0 provisioning, controller and container monitoring, enhanced Solid-State Drive (SSD) and interface telemetry, and orchestration changes for multi-tenant Artificial Intelligence (AI) infrastructure. The update matters to enterprise IT and security leaders managing hybrid network fabrics, operations visibility, and tenant segmentation.

Research Overview

The vendor describes ONES 4.2 as an update to the Open Networking Enterprise Suite that unifies Graphics Processing Unit (GPU), front-end, and storage networking management in a single platform. The post frames the release around tighter automation, expanded observability, and support for hybrid fabric deployments that include NVIDIA SpectrumX and SONiC-based environments.

The blog also outlines changes intended to improve tenant isolation and programmability in AI workloads, including multi-tenant compute sharing on the same physical server. It further highlights interface and storage telemetry expansions for operational visibility.

Key Findings

ONES 4.2 adds Day-0 automation for onboarding by taking a Monitoring-as-Code (MaC) address and serial number. The workflow includes Dynamic Host Configuration Protocol (DHCP) configuration, device IP allocation, and dynamically generated Zero Touch Provisioning (ZTP) scripts for steps such as Network Optimization Suite (NOS) image installation, configuration restoration, and deployment of ONES telemetry and configuration agents.

The release introduces time-series monitoring for controller Central Processing Unit (CPU), memory, and disk utilization and adds rule engine alerting. It also adds container-level (Docker) monitoring and alerting for individual ONES services running within the controller.

Technical Breakdown

For licensing, ONES 4.2 provides a flexible model that lets customers select licenses based on Operating System (OS) type, switch port speeds, and separate fabric design licenses. The licensing dashboard is described as showing all licenses, used versus available seats, and options to reassign devices between licenses.

Storage observability expands with additional SSD-level telemetry metrics, including erase fail count, Serial ATA (SATA) CRC errors, power cycles, unsafe shutdown events, unexpected power loss, late bad blocks, total bad blocks, spare block availability, SSD wear level (health index), total host writes and reads, and power-on hours. Interface monitoring adds time-based link flap history to the interface down widget.

AI Infrastructure Orchestration Enhancements

The blog describes changes to GPU infrastructure orchestration by moving from physical server-level tenant binding toward multi-tenant compute Virtual Machine (VM) sharing on the same server and fine-grained GPU allocation at the individual GPU level. It states that the orchestration prevents cross-tenant interference and enforces that a single GPU is never shared across tenants.

For front-end to storage connectivity, ONES 4.2 is described as adding API-driven workflows that manage connectivity without explicit peering constructs, while using tenant segmentation based on Virtual Extensible Local Area Network (LAN) (VXLAN) or Virtual LAN (VLAN) plus Virtual Routing and Forwarding (VRF). The post also calls out unified behavior across platforms and a hybrid telemetry model that combines NVUE with gNMI for streaming telemetry and detailed visibility.

Overall, ONES 4.2 combines Day-0 bootstrapping, expanded controller and container monitoring, added SSD and interface telemetry, flexible licensing, and orchestration changes for multi-tenant GPU allocation and programmable front-end to storage connectivity. This “Blog Signals brief” is a fact-based summary of the vendor blog.