Aviz Networks details ONES, ASN, and Network Copilot for AI factories
Aviz Networks, working with Red Hat, outlines an approach for managing AI factory networks by pairing an orchestration layer with packet-level observability and AI-assisted operations on Red Hat Enterprise Linux and OpenShift. The update targets multi-tenancy, streaming workflows, and real-time visibility needs that emerge when workload automation and network operations run separately.
Research Overview
The blog describes AI factory environments as combining highly dynamic workloads and distributed systems with ongoing requirements for standard enterprise network infrastructure. It frames the central issue as a persistent gap between workload orchestration and network operations as environments expand.
It lists three recurring requirements: multi-tenancy management across applications and infrastructure, support for streaming operations from Day-0 through Day-2 with telemetry for observability, and real-time visibility into network and application behavior.
Key Findings
The post argues that workload orchestration and network operations are frequently handled by different teams and tools, which complicates tenant isolation and troubleshooting. It says the gap makes it harder to maintain consistent control as scale increases and network and application traffic becomes harder to monitor.
It presents a coordinated approach intended to extend orchestration into the network layer while keeping the operational model aligned with enterprise Linux and OpenShift environments.
Technical Breakdown
The blog centers on Aviz ONES as the network orchestration layer for AI fabrics, positioned to orchestrate network infrastructure, support multi-tenancy and tenant isolation, and simplify operations across the infrastructure lifecycle from Day-0 to Day-2. It says ONES is designed to align network operation with the orchestration model used for workloads.
For observability, Aviz Service Node (ASN) is described as providing packet-level intelligence, including deep packet inspection for application identification, packet deduplication, GPRS tunneling protocol (GTP) correlation, metadata enrichment, and packet capture for troubleshooting. For operations, Aviz Network Copilot is described as using AI capabilities to identify issues faster by reducing mean time to repair (MTTR), correlate signals across infrastructure layers, and perform root cause analysis (RCA).
Operational Impact
The blog states that Aviz ONES, ASN, and Network Copilot run as software on Red Hat Enterprise Linux. It says this supports deployment across connected, disconnected, and air-gapped setups, with a consistent operational model across on-prem, cloud, and edge environments.
It also includes an operations workflow theme in which telemetry is positioned as real-time rather than periodic reporting, citing the need to catch network problems during fast-moving AI training and development cycles. The blog links the approach to more coordinated handling of infrastructure, networking, and operations in a unified system.
Aviz Networks and Red Hat describe a network-centric orchestration and observability stack for AI factory environments, centered on ONES for network orchestration, ASN for packet-level visibility, and Network Copilot for AI-assisted operations, all running on Red Hat Enterprise Linux and OpenShift. This Blog Signals brief is a fact-based summary of the vendor blog.