Community
Community

Exciting Announcement! In celebration of launching our AI Certification, we’re thrilled to offer a 50% discount exclusively. Seize this unique chance—don’t let it slip by!

Learn more

Category: Fabric Test Automation Suite

Your blog category

Categories
Fabric Test Automation Suite

Why We’re Hosting the Observability Interop — And Why It Matters to You

Legacy observability stacks — packet brokers, service nodes, and closed analytics platforms — weren’t built for today’s scale, AI-driven agility, or cost-sensitive infrastructure teams. If you’re running visibility infrastructure today, chances are you’ve hit some or all of these roadblocks:

  • Budget spikes every time you refresh hardware
  • Rigid, hardware-bound architectures that restrict flexibility
  • Vendor-controlled dashboards that limit automation and data ownership
  • Sky-high per-port pricing with little ROI clarity
  • Outdated power and rack footprints that ignore modern efficiency needs

We believe observability needs a reset — not another patch.

That’s why we’re hosting the Deep Network Observability Interop:
 A community-driven showcase of what’s possible when observability becomes modular, software-first, AI-ready, and vendor-agnostic.

Here’s how Vishal Shukla, CEO of Aviz Networks, frames it:

This is the inflection point. Infrastructure teams deserve choices, transparency, and automation that scale. The Interop proves that you can move away from rigid stacks without compromising on performance, insights, or control.

Vishal Shukla, CEO of Aviz Networks,

See real configs. Real results. No sales pitch — just better architecture.

Join the Interop
Categories
Fabric Test Automation Suite

Building a Smarter, Standardized Retail Network

The Challenge: Scaling Without Lock-In

Managing a vast global retail footprint demands a network that’s both flexible and predictable. For a major retailer operating thousands of stores and multiple data centers, the goal was ambitious but clear: standardize network operations end-to-end, avoid vendor lock-in, and reduce total cost of ownership.

The Open Networking Advantage

To achieve this, the company embraced open networking for retail chains, deploying a vendor-neutral, disaggregated networking approach built on Enterprise SONiC deployments. Backed by robust NOS lifecycle automation and support from Aviz Networks, they unified their data center and in-store networks under a single, adaptable architecture compatible with any leading ASIC or switch vendor.

How Aviz Delivered at Scale

From proof-of-concept to large-scale rollout, Aviz provided certified SONiC distributions, CI/CD pipelines for automated updates, and enterprise-grade support — ensuring smooth retail infrastructure modernization and consistent network performance across every location.

Results That Matter: Lower Costs, Greater Flexibility

The results were decisive: upgrade costs dropped by 50%, budget surprises disappeared, and the network gained the freedom to scale on the company’s terms. By leveraging open networking for retail chains, the retailer now adapts quickly to supply chain shifts and new market demands without costly proprietary limitations.

Why It’s Time for Open, Standardized Retail Networks

For retailers planning the next decade of growth, Enterprise SONiC deployments and disaggregated networking deliver the control, savings, and agility that proprietary stacks can’t match.

Explore the full case study to learn how enterprise SONiC deployments are helping top retailers standardize, reduce TCO, and escape vendor lock-in.

Read Case Study
Categories
Fabric Test Automation Suite

Building Efficient Networks for AI: Cisco and Aviz Networks Bootcamp Recap

TL;DR

  • How to design lossless, high-performance fabrics for AI/LLM workloads
  • Cisco 8000 series with Silicon One + SONiC explained
  • Techniques for RDMA over Converged Ethernet (RoCE v2), Priority Flow Control (PFC), and Explicit Congestion Notification (ECN)
  • How Aviz ONES delivers multi-vendor observability and real-time congestion insights
  • Validated leaf-spine topologies and blueprints for scalable AI clusters

As data centers evolve from CPU-centric to GPU-centric architectures, building networks that can handle massive AI workloads has become mission-critical. At the recent Cisco and Aviz Networks joint bootcamp, attendees got an inside look at how to design modern, lossless, high-performance networks for AI and LLM workloads.

Figure 1: Solving AI Challenges with SONiC

Satish Surani, Director of Product Management at Cisco, showcased how the Cisco 8000 series, powered by Silicon One and SONiC, delivers a flexible, open networking stack designed for today’s demanding AI clusters. Key topics included the importance of RDMA over Converged Ethernet (RoCE v2), Priority Flow Control (PFC), and Explicit Congestion Notification (ECN) to maintain low latency and lossless traffic under heavy GPU workloads.

Figure 2: Cisco 8000 Networking Topology Diagram

Aviz demonstrated how its Open Networking Enterprise Suite (ONES) and real-time telemetry can monitor and optimize congestion points, validate QoS policies, and provide deep observability across multi-vendor environments.

The session also covered best practices for building scalable, non-blocking leaf-spine topologies and how Cisco and Aviz jointly deliver validated blueprints for large-scale AI infrastructure.

As AI clusters grow, open, programmable fabrics with robust congestion management are no longer optional — they’re foundational for faster job completion and predictable performance.

Learn More: Explore Aviz ONES

Watch the full bootcamp on-demand

View the bootcamp here
1. What was the main focus of this joint bootcamp?

This bootcamp explored how to design AI-ready data centers using Cisco 8000 switches running SONiC NOS — and how Aviz ONES delivers observability, congestion management, and telemetry for lossless, high-bandwidth AI workloads.

 AI clusters require:

  • High bandwidth
  • Low latency
  • Zero packet loss for RDMA flows

Even slight congestion can increase training job completion time. Techniques like PFC and ECN keep traffic lossless and predictable.

Remote Direct Memory Access (RDMA) allows data transfer directly between server memory without CPU involvement, which:

  • Reduces latency
  • Increases throughput
  • Optimizes GPU-centric workloads

ROCE v2 (RDMA over Converged Ethernet) was discussed as a practical method for enabling RDMA on Ethernet networks.

Explicit Congestion Notification (ECN) marks packets experiencing congestion instead of dropping them outright. End devices then get feedback to slow down traffic rates, preventing packet loss and ensuring consistent AI training performance.

 PFC provides lossless service for specific traffic classes. In AI networks, PFC:

  • Guarantees no packet drops for RDMA traffic
  • Enables mixed traffic without sacrificing critical flows
  • Works in tandem with ECN for robust congestion management.

Cisco’s Silicon One G200 chip powers switches with:

  • Ultra-low latency
  • High radix (51.2 Tbps capacity)
  • Fully shared packet buffers for burst absorption

These features enable massive GPU clusters with efficient load balancing and job completion time optimization.

SONiC provides an open, modular network operating system that customers can fully customize. Cisco 8000 switches run SONiC with enterprise support, ensuring open standards, rapid feature updates, and community-driven innovation.

ONES acts as an agentless observability and orchestration layer:

  • Collects multi-vendor telemetry
  • Monitors PFC, ECN, and RDMA health
  • Visualizes real-time AI traffic versus non-AI traffic
  • Automates troubleshooting across leaf-spine and super-spine topologies.

The live demo featured:

  • Cisco 8100 switches configured for PFC and ECN
  • Factory topology design with non-blocking bandwidth
  • Real-time visibility of congestion points and queue health via Aviz ONES
     This demonstrated how to keep AI traffic lossless and optimize cluster efficiency.

 ONES is designed for openness:

  • It supports SONiC, Cisco XR, Arista EOS, and more.
  • Uses standard APIs and collectors — no heavy agents.
  • Offers a unified observability dashboard for hybrid and AI fabrics, regardless of hardware vendor.
Categories
Fabric Test Automation Suite

Upgrading Your Network? Why More Operators Are Migrating to SONiC

As networks grow more complex and budgets tighten, many operators are rethinking traditional, proprietary networking stacks. In our recent bootcamp, Aviz Networks shared how open networking with SONiC delivers cost savings, flexibility, and modern automation without vendor lock-in.

Why SONiC?

SONiC (Software for Open Networking in the Cloud) disaggregates your network operating system from hardware — just like Linux did for compute. This means you can run SONiC on white-box switches from vendors like Edgecore, Celestica, Supermicro, and silicon from Broadcom, Marvell, or Cisco Silicon One.

Key benefits:

  • Freedom of choice: Mix and match hardware and silicon.
  • Greater control: Modular microservices let you patch or upgrade independently.
  • Lower TCO: Operators see 40–60% cost savings versus proprietary NOS licenses.
  • Built-in quality: Aviz’s Fabric Test Automation Suite (FTAS) validates your deployment with end-to-end regression and scale tests.

During the session, the team walked through real migration scenarios — from CLI-based per-device migration to fabric-wide deployment using Aviz ONES. They showed how validated templates, YAML-based intent, and automation make Sonic adoption practical and low-risk.

"Just like Linux runs on any computer, your network operating system should run on any hardware — that’s the vision we’re bringing to open networking."

Kuram Khani, Aviz Networks

Watch the full bootcamp demo to see how operators migrate to SONiC step-by-step

Learn how Aviz makes SONiC migration easy and cost-effective

Explore SONiC with Aviz
Categories
Fabric Test Automation Suite

Smarter Network Visibility with Aviz Service Node

As telecom operators scale 4G and 5G networks, legacy monitoring tools often struggle to handle growing traffic and multiple technologies. In a recent bootcamp, Aviz Networks showcased how its Aviz Service Node (ASN) modernizes network visibility with a flexible, software-based approach.

What is ASN?

 ASN is a high-performance, DPDK-powered packet processing appliance that runs on standard x86 servers. Integrated with the Open Packet Broker (OPB) and Aviz’s Network Co-Pilot, it forms a complete, vendor-neutral visibility stack.

Key benefits:

  • Unified monitoring: Covers 4G, 5G NSA/SA, data centers, and enterprise sites.
  • Deep Packet Inspection: Identifies encrypted apps like WhatsApp and Facebook using multiple techniques.
  • Per-subscriber KPIs: Correlates control and user plane flows for precise throughput, latency, and session metrics.
  • Packet de-duplication: Removes duplicate packets from multiple taps, saving bandwidth and tool costs.
  • Smart load balancing: Distributes subscriber sessions across tools to prevent overload.
  • Open data export: Streams enriched metadata to Elasticsearch, Kibana, Kafka, or any data lake.

"Operators need a vendor-neutral way to scale packet processing on commodity servers. ASN delivers 50% cost savings by removing proprietary hardware and licensing fees. It gives customers choice, control, and significant savings — the three C’s that define our value"

Ilanchezhian Raman, Senior Solution Architect at Aviz

All ASN nodes are managed through Flow Vision, Aviz’s centralized UI for orchestration and telemetry.

Watch the full bootcamp demo to see how operators replace costly proprietary probes with flexible, software-based packet processing.

Discover how ASN improves visibility, reduces tool sprawl, and saves up to 50%
Explore Aviz Service Node
Categories
Fabric Test Automation Suite

5 Reasons to Start Your SONiC Journey at the OCP + Aviz ONE Center

The industry is converging on open networking—and SONiC is leading the way. But getting started with SONiC in a real-world, multi-vendor environment requires more than just code. It requires infrastructure, validation, community support, and forward-thinking automation. That’s why the OCP + Aviz ONE Center, hosted in Fremont California, has quickly become the default launchpad for serious SONiC adopters.


Here are five reasons why your SONiC journey should begin here.

1. The First and Only Official End-to-End SONiC Experience Center by OCP

The OCP Experience Center, hosted by Aviz, is not just a lab—it’s the only officially recognized end-to-end SONiC experience center backed by the Open Compute Project. It supports the full SONiC lifecycle from:

  • Pre-deployment PoC
  • Vendor comparison and integration
  • CVE vulnerability scans and posture analysis
  • Operational validation
  • Post-deployment automation and support

All while staying 100% vendor-neutral and community aligned.

2. Trusted by Top F100 for Real-World SONiC Lifecycle Testing

This isn’t a sandbox—it’s the proving ground where Fortune 100 companies validate and optimize their SONiC deployments across diverse ASICs and platforms. These enterprises use the center to simulate real traffic, test upgrade scenarios, validate CVE fixes, and replicate day-2 operations—before ever touching production.

With built-in CVE analysis, the lab continuously evaluates known vulnerabilities across the stack, giving teams a clear security posture and remediation strategy. Curious how the OCP + Aviz ONE Center supports these top-tier customers? Ping us here to open a conversation with our team.

3. PlugFest Proven + Validated Configs Ready to Deploy

The SONiC PlugFest hosted at the ONE Center proved that this isn’t theoretical. Major hardware vendors, ASIC platforms, and operators came together to test real interoperability and day-2 scenarios in a production-like setting. It tested use cases like IPCLOS, EVPN, RoCE (for AI Fabrics), MC-LAG and Edge use cases.

Thanks to OVD.ai, you also get ready-to-use, 100% validated SONiC YAML configurations—customizable templates you can deploy on your own network with confidence. No guesswork. No misconfigurations.

4. Multi-Vendor, Multi-ASIC, and end to end SONiC Support with ONES-Enabled automation

Aviz’s ONE Center provides a unique setup that includes Broadcom, NVIDIA, Intel, and other ASICs across multiple whitebox vendors—tested daily with ONES and FTAS. This ensures your team can test anything—from basic feature compatibility to regression across software versions, at a pace that matches cloud-native DevOps environments. The quality comes with ONES integrated, so when you move to deployment stages you are well prepared for it:

  • Vendor-agnostic orchestration for complex features (BGP, EVPN, etc.)
  • 200+ telemetry metrics, global multi-site observability.
  • Seamless CI/CD integration via FTAS for regression testing

5. SONiC Is Just the Beginning – AI-Powered Automation Is the Future

SONiC gives you vendor freedom. But what’s next?

Standardizing your automation layer is the natural next step—and that’s where Aviz’s Network Co-Pilot comes in. Once SONiC is up and running, Network Co-Pilot integrates with it to power:

  • Closed-loop troubleshooting
  • AI-driven anomaly detection
  • Multi-vendor configuration workflows
  • Security audits
  • Auto-generated intents and YAMLs
  • Integration with your own workflows and playbooks

And best of all—it’s LLM-ready, letting you use the best open-source AI models to turn your SONiC deployments into AI-powered network infrastructure.

Final Word

There’s no better place to begin your SONiC journey than the OCP + Aviz ONE Center. Whether you’re evaluating open networking, comparing vendors, validating CVEs, or deploying AI-based automation, this is the most complete, future-ready SONiC ecosystem available today.

Ready to experience it yourself?

Book your time at the ONE Center.
Categories
Fabric Test Automation Suite

Accelerating AI Networking: NVIDIA Spectrum-X and Aviz ONES Integration

As AI workloads continue to scale, enterprises are quickly realizing that traditional networking isn’t built for the demands of modern, distributed GPU clusters. To address this, NVIDIA and Aviz Networks hosted a joint bootcamp showcasing the deep integration between NVIDIA Spectrum-X — the industry’s first Ethernet fabric optimized for AI — and the Aviz Open Networking Enterprise Suite (ONES), purpose-built for AI infrastructure orchestration and observability.

"AI had its iPhone moment with ChatGPT. Suddenly, enterprises everywhere wanted to deploy generative AI at scale — but Ethernet couldn’t keep up."

Dave Isles, Senior Director of AI Networking, NVIDIA

How is Ethernet evolving for AI clusters?

From InfiniBand to Ethernet: The AI Networking Shift

Enter Spectrum-X. NVIDIA took key capabilities from InfiniBand and extended them to Ethernet — enabling RDMA, adaptive routing, and congestion control, all with the governance and familiarity enterprises demand from Ethernet environments.

"The larger the AI cluster, the bigger the impact the network has on performance, Spectrum-X gives enterprises a purpose-built Ethernet fabric to unlock full GPU performance.

David Iles, Senior Director at NVIDIA Corporation
Figure 1: NVIDIA Spectrum-X Full Stack + Aviz ONES Network Orchestration Diagram

What is Spectrum-X RA 1.3.0 and how is it validated?

Figure 2: Spectrum-X Network Orchestration Ecosystem

Spectrum-X RA 1.3.0: Validated at Supercomputer Scale

Aranga Madipuri, Product Manager at NVIDIA, detailed the Spectrum-X Reference Architecture (RA 1.3.0), tested on real-world supercomputers like Israel-1. The RA offers a prescriptive blueprint combining SONiC/Cumulus NOS, NetQ telemetry, NVIDIA AIR digital twin simulation, and BlueField-accelerated switching — ensuring performance, reliability, and reproducibility for massive AI clusters.

How does Aviz ONES support Spectrum-X deployments?

Figure 3: ONES for Spectrum-X Monitoring & NetOps

ONES by Aviz: Purpose-Built for Multi-Tenant AI Infrastructure

Aviz CTO Chit Perumal and Principal Engineer Kasi Nath demonstrated how ONES seamlessly integrates with Spectrum-X RA 1.3.0. ONES delivers:

  • Day 0–2 automation: Declarative fabric design, NVIDIA AIR simulation, automated switch and host configuration using RA-aligned templates.
  • Multi-tenancy orchestration: EVPN and VRF-based segmentation, GPU-aware resource provisioning, and policy-driven isolation across workloads.
  • Telemetry and alerting: Agentless, real-time visibility from switches, hosts, and GPUs — plus built-in alerting integrated with Slack, ServiceNow, and Zabbix.
  • Lifecycle operations: Config drift detection, backup/restore, structured RMA workflows, and topology-aware fabric comparison.

"We wanted customers to scale GPU clusters effortlessly while maintaining network visibility and operational simplicity — ONES makes that possible."

Chid Perumal, CTO, Aviz Networks

What was covered in the live demo?

Real-World Topologies, Live Demo

The session closed with a detailed demo covering:

  • Automated orchestration of a two-SU Spectrum-X fabric
  • Tenant creation and GPU assignment
  • Policy-driven isolation validation
  • Real-time monitoring dashboards and anomaly detection
  • Full config comparison and RMA workflows

Explore More

  • Watch the full bootcamp recording
  • Learn about Aviz ONES

Whether you’re building a private AI cloud or launching GPU-as-a-Service, the NVIDIA + Aviz stack gives you the tools to scale with confidence — and visibility.

Frequently Asked Questions

1. What is NVIDIA Spectrum-X and how is it optimized for AI workloads?

Spectrum-X is the first Ethernet fabric designed specifically for AI clusters. It extends InfiniBand’s low-latency, lossless transport to Ethernet, delivering RDMA, adaptive routing, and congestion control — all with familiar enterprise Ethernet governance.

Traditional Ethernet often struggles with network congestion and packet loss during large-scale GPU training. Spectrum-X solves this by:

  • Enabling RDMA over Converged Ethernet (RoCE) for direct GPU communication.
  • Using adaptive routing to bypass congestion.
  • Delivering consistent throughput across thousands of GPUs.

It’s a validated deployment blueprint tested on real supercomputers like Israel-1. RA 1.3.0 combines:

  • Open NOS (SONiC/Cumulus).
  • NetQ telemetry.
  • NVIDIA AIR for digital twin simulation.
  • BlueField DPUs for hardware acceleration.

This ensures predictable performance and easier scaling of AI networks.

ONES is a software layer for orchestration and observability. It connects directly with Spectrum-X fabrics to automate deployment, manage multi-tenant AI workloads, and deliver agentless, real-time telemetry — reducing operational complexity.

  • Declarative fabric design: Define your network layout upfront.
  • NVIDIA AIR simulation: Test configurations in a digital twin before deployment.
  • Zero-touch provisioning: Automatically configures switches and hosts using Spectrum-X RA templates.

ONES uses EVPN and VRF-based segmentation to isolate traffic between tenants. It also provisions GPU resources intelligently and applies policies to guarantee secure workload separation.

ONES leverages built-in telemetry from the NOS, hosts, and GPUs. It integrates with tools like Slack, ServiceNow, and Zabbix for automated alerting — without installing third-party agents that consume resources.

  • Orchestration of a two-switch Spectrum-X fabric
  • Tenant creation and GPU assignment
  • Policy validation for isolation
  • Real-time monitoring dashboards and anomaly detection
  • Config comparison and structured RMA workflows.

Organizations running large AI training clusters, GPU-as-a-Service providers, and any enterprise building private AI clouds that require high performance, robust isolation, and full-stack observability.

Watch the bootcamp recording for a full walkthrough and explore Aviz ONES for detailed docs, case studies, and a deeper look at deployment best practices.

Categories
Fabric Test Automation Suite

Unlock AI-Driven NetOps with Network Copilot

In today’s complex multi-vendor environments, network teams spend countless hours monitoring, troubleshooting, and documenting — often juggling disparate tools and repetitive tasks. Aviz Networks’ recent bootcamp showcased how Network Copilot, an AI-powered networking assistant, redefines this workflow.

Figure 1: Aviz Network Copilot

Built as a private AI platform, Network Copilot securely runs on-premises, keeping your data within your environment. It uses an agentic architecture to automate routine tasks like upgrade compliance checks, security audits, and hardware lifecycle tracking — freeing engineers to focus on higher-value work.

Figure 2: Agentic Architecture

A standout feature is its retrieval-augmented generation (RAG) pipeline: real-time telemetry is enriched with up-to-date knowledge bases to deliver precise, context-aware answers. Users interact naturally via a unified chat interface or familiar tools like Slack, asking questions in plain language and receiving actionable insights instantly.
Teams can customize agents, templates, and even choose preferred LLMs for different use cases — ensuring flexibility as networks evolve.

Ready to see it in action? Explore how Network Copilot helps engineers work smarter, not harder.

Explore Network Copilot

Watch the full bootcamp on-demand

Explore Network Copilot

Frequently Asked Questions

1. What is Network Copilot?

Network Copilot is an AI-powered networking assistant that automates repetitive NetOps tasks and provides real-time, context-rich insights — all while keeping data private.

  • Runs fully on-premises
  • No public cloud training on your data
  • Air-gapped or single-server deployments possible
  • Upgrade compliance checks
  • Security audits
  • Incident response
  • Troubleshooting
  • Documentation and reporting

Unlike general LLM chatbots, Copilot is designed specifically for networking. It combines live telemetry with up-to-date knowledge, uses an agentic system for accurate answers, and fits any multi-vendor environment.

Yes!

  • Integrates with Slack and ticketing systems
  • Uses standard protocols (SNMP, gRPC, NetFlow, etc.)
  • Connectors for SONiC, Arista, Cisco, Splunk, NetBox, and more

Absolutely.

  • Create custom agents for new tasks
  • Use your own compliance templates
  • Pick preferred LLMs per use case

The retrieval-augmented generation (RAG) pipeline combines real-time telemetry with curated knowledge bases — ensuring answers are up-to-date, relevant, and free from typical LLM hallucinations.

  • Deploy on-premises servers with NVIDIA GPUs (e.g., A100)
  • Runs fully air-gapped if needed
  • Scales horizontally with containers or Kubernetes

Network engineers, operations teams, security admins — anyone managing complex, multi-vendor network environments who wants faster insights and less manual work.

Getting started is simple:

  • Request a trial or test drive in your own lab.
  • Connect it to a small multi-vendor network to see real results.
  • Use our quick-start guides and templates to deploy in hours — not weeks.
     For details, visit the Network Copilot page.
Categories
Fabric Test Automation Suite

How AI is Reshaping Networking—and Why SONiC is Leading the Charge

A Conversation with Vishal Shukla, CEO of Aviz Networks

In a recent conversation, Alan Weckel of the 650 Group sat down with Vishal Shukla, Co-founder and CEO of Aviz Networks, to discuss the evolution of open networking, the growing maturity of the SONiC ecosystem, and how artificial intelligence is redefining both network infrastructure and operations. Here are the key insights from that conversation.

What Is Aviz Networks and Why Does It Matter in AI Networking?

"We are creating a network software stack, end to end, in two dimensions — Networks for AI and AI for Networks. That means building networks that support AI workloads, and using AI to make networks smarter."

Vishal Shukla, CEO, Aviz Networks

Aviz Networks builds a full-stack, vendor-agnostic platform for open networking and AI-powered automation. At the center is ONES (Open Networking Stack), which provides a unified control and observability layer across NOS platforms like SONiC and NVIDIA Spectrum-X.
This approach gives enterprises the flexibility to scale their infrastructure while maintaining operational consistency—whether they’re deploying AI workloads at the edge or automating compliance in the data center.

How Is Aviz Executing on Its Vision for Open Networking?

Over the past two years, Aviz has transitioned from vision to deployment by building a fully integrated, AI-ready networking stack:

  • 2023–2024: Standardized SONiC integration for vendor-agnostic networks
  • Late 2024: Introduced deep visibility and telemetry tools for real-time observability
  • 2025: Launched Network Copilot, an AI assistant that enables proactive, data-driven network automation

Today, Aviz is deploying its stack across 15 verticals, including healthcare, telco, e-commerce, and retail—proving that open networking and AI networking are enterprise-ready.

Why Is SONiC Adoption Accelerating in Retail and Edge Use Cases?

Retail has become a standout use case for SONiC adoption, especially at the edge. Originally designed for hyperscale data centers, SONiC has now been extended to store-level and branch environments through the Aviz-led PENCE initiative (POE Edge Networks with SONiC).

Edge networks in retail demand:

  • High availability with minimal complexity
  • Low cost of operation
  • Real-time visibility into remote locations

SONiC’s lightweight, scalable architecture—combined with Aviz’s enhancements—makes it an ideal fit for these needs.

How Has SONiC’s Evolution Prepared It for Enterprise-Scale Network Automation?

SONiC has evolved into a production-grade solution for network operators across industries. According to Vishal, this evolution spans three dimensions:

  • Protocols: Support for IP CLOS, EVPN, VXLAN, and edge features now covers 80% of enterprise use cases.
  • Enterprise Readiness: Faster patching and microservices architecture enable quick responses to security vulnerabilities (e.g., OpenSSL CVE fix within 3 weeks).
  • Deployment Maturity: Proven use across healthcare, telco, and retail verticals—with or without Aviz—validates SONiC’s scalability and flexibility.

This evolution positions SONiC as a core enabler of AI networking and next-gen network infrastructure.

How Is the SONiC Ecosystem Growing Around Open Networking?

As enterprise needs grow more complex, the SONiC ecosystem is expanding to include:

  • Orchestration and automation layers
  • Real-time observability and SLA monitoring
  • Integration with SecOps and NetOps tools
  • 24/7 enterprise support
  • Security and compliance frameworks

Aviz is actively building the controller layer around SONiC through its ONES platform—tying together telemetry, automation, and AI to provide a turnkey solution for open networking at scale.

What Role Does SONiC and ONES Play in Networks for AI?

As AI workloads scale, enterprises need flexible, high-bandwidth fabrics. Two deployment models are emerging:

  • NVIDIA Spectrum-X for vertically integrated AI environments
  • SONiC-based fabrics for open, customizable deployments

Aviz’s ONES platform supports both approaches—providing a consistent operational interface across front-end and back-end networks. In Vishal’s words:

“In this world where the infrastructure is changing very fast... you need to standardize your operating layer.”

Vishal Shukla, CEO, Aviz Networks

By decoupling the control plane from the underlying hardware, Aviz allows customers to support 400G, 800G, and 1.6T environments—future-proofing the network for accelerated AI innovation.

How Does AI for Networks Drive Real-World Network Automation?

AI for Networks is all about applying intelligence to daily operations. Aviz Network Copilot is designed to automate routine NetOps workflows through vendor-agnostic, privacy-preserving AI:

  • Compliance Checks – Continuously monitor configuration, security, and operational compliance
  • Audit Report Generation – Use generative AI to create structured documentation
  • Playbook Standardization – Ensure team-wide consistency in automation scripts
  • Toolchain Integration – Connect to JIRA, ServiceNow, Zabbix, and other NetOps tools

By embedding network automation into everyday workflows, Copilot reduces manual effort and enables teams to focus on strategic tasks.

“It’s like a human network engineer—AI must be vendor-agnostic and understand the full stack.”

Vishal Shukla, CEO, Aviz Networks

What’s an Example of Network Copilot in Action?

In one customer deployment, each engineer was writing their own Ansible playbooks—resulting in fragmented automation. With Network Copilot, Aviz enabled:

  • Standardization of scripting formats
  • Role-based customization
  • Automated playbook generation with AI-assisted validation

This use case demonstrates how AI can enable repeatable, secure, and scalable network automation.

What’s Next for AI Networking in 2025?

Aviz is doubling down on two fronts:

1. Scaling SONiC and Open Networking for AI Workloads

  • Broaden ONES + SONiC + Spectrum-X deployments across all 15 verticals
  • Continue expanding SONiC’s feature set for edge and AI fabric use cases

2. Defining AI’s Role in Network Automation

  • Identify 10–20 repeatable use cases for Network Copilot
  • Validate these use cases across real-world customer environments
  • Refine agent-based orchestration for NetOps and SecOps workflows

The end goal? A new standard for how enterprise networks are designed, operated, and automated.

Explore Aviz ONES and Network Copilot

See how Aviz is redefining what’s possible with open networking and AI-powered automation.
  • Learn more about ONES
  • Discover Network Copilot

Frequently Asked Questions

AI Networking, SONiC Adoption, Open Networking, and Network Automation

1. What is SONiC, and how does it support AI Networking?

SONiC (Software for Open Networking in the Cloud) is an open-source network operating system originally developed by Microsoft. It’s now widely used by enterprises to build vendor-agnostic, high-performance network infrastructure.
SONiC enables flexibility, scalability, and cost savings—making it ideal for supporting AI workloads across data centers and edge environments. Its modular architecture allows integration with platforms like Aviz ONES, providing full-stack support and observability.

Explore Aviz’s SONiC-based Open Packet Broker

Network Copilot is an AI-driven assistant that automates everyday NetOps tasks. It continuously monitors network health, flags compliance issues, and generates audit reports using generative AI.
Unlike traditional tools, it is vendor-agnostic, integrates with JIRA, ServiceNow, and Ansible, and can operate in on-prem environments for full data privacy. It’s built to scale across multi-vendor environments, making it ideal for enterprises looking to adopt intelligent, secure network automation.

Learn more about Network Copilot

ONES is a vendor-neutral operating layer that connects SONiC, NVIDIA Spectrum-X, and third-party observability and orchestration tools. It supports real-time telemetry, SLA enforcement, and multi-domain visibility, all while preserving the open-source nature of SONiC.
What makes it unique is its ability to serve as a single control point for Networks for AI—handling both backend AI fabric and frontend application layers.

Discover ONES: Open Networking Stack

SONiC has significantly matured across protocols, enterprise readiness, and ecosystem support. It now supports IP CLOS, EVPN, VXLAN, and edge features—covering 80%+ of enterprise use cases.
Security responsiveness has improved thanks to its microservices model, and deployments span industries like healthcare, retail, telco, and e-commerce. With vendors like Aviz providing support, observability, and controller layers, SONiC is ready for mainstream enterprise use.

AI improves operational efficiency by offloading mundane, repetitive tasks—like config validation, compliance checks, and documentation generation. Tools like Network Copilot act as co-pilots, helping network engineers troubleshoot faster, enforce standards, and reduce human error.
The key? It’s agentic—meaning it doesn’t just analyze data, but also triggers workflows, integrates with automation pipelines, and adapts to unique infrastructure setups.

Categories
Fabric Test Automation Suite

Unlocking Deep Network Observability: Lessons from the Field

When it comes to modern networking, what you can’t see can hurt you. At Aviz Networks, we make the invisible visible—before problems escalate. In a recent bootcamp, we explored how deep observability empowers teams to reduce troubleshooting time, achieve compliance, and build more resilient infrastructure. Through a customer case study, a partner perspective, and a full-stack demo, we showed how organizations are moving from reactive to proactive network operations.

What Happens When Visibility Fails? A Real-World Crisis Unpacked

Frank Vincentelli, CTO of managed service provider Integrated IT, shared a powerful story. After upgrading the infrastructure for a mid-market customer, his team faced a major network outage with no clear root cause. Multiple vendors pointed fingers. Logs were incomplete. Visibility was limited.
With Aviz’s platform—including the Open Packet Broker (OPB) and Aviz Service Nodes (ASN)—Frank’s team was able to rule out false leads and identify the actual source of failure.
Because Aviz embeds lightweight agents within switches, telemetry continued even when the network was down. Once connectivity was restored, historical data flowed in, providing the “X-ray vision” they needed.

“The product is excellent—but it’s the people behind it that made the difference, with Aviz, we transitioned from a dysfunctional vendor relationship to one that’s transparent and proactive.”

Frank Vincentelli, CTO of managed service provider Integrated IT

Why Openness Matters: A Partner’s View from the Field

Joel Winland, Field CTO at Source Code, discussed their strategic alignment with Aviz. Source Code, a global system integrator with nearly $1 billion in revenue, builds both edge and cloud-scale infrastructure. What they needed was a software partner as open and flexible as they are.
Aviz fit that need. Built on open-source principles—including full support for SONiC—Aviz allows customers to choose the hardware, NOS, and deployment model that best fits their architecture. No vendor lock-in. No black boxes. Just choice and control.

Comparison table highlighting Aviz’s advantages across open-source support, AI-driven insights, vendor neutrality, observability depth, and cost efficiency.

How Does Aviz Deliver End-to-End Observability? A Look Inside the Stack

The Aviz platform is composed of three tightly integrated components:

  • Open Packet Broker (OPB): A high-performance, white-box-based traffic management fabric. OPB handles traffic filtering, replication, aggregation, slicing, labeling, and tunneling. It supports granular filtering from L2–L4 and enables advanced capabilities such as GTP header parsing for 5G environments and symmetric load balancing.
  • Aviz Service Nodes (ASN): These x86-based software appliances perform deep packet inspection, metadata extraction, KPI generation, application identification, and deduplication. ASN works for both telco (4G/5G) and Enterprise, , providing Application and per-user/session visibility and reducing tool overhead by filtering out redundant, high, and low-risk traffic.
  • Network CoPilot (NCP): An AI-powered, on-premises assistant built on open-source LLMs. NCP allows users to query real-time network data using natural language. It’s RAG-enabled, privacy-safe (runs locally), and avoids hallucinations by relying strictly on verified telemetry.

All of this is managed through Flow Vision, a centralized GUI and single pane of Glass that simplifies traffic orchestration, rule configuration, tunnel management, and health monitoring across OPB and ASN nodes.

End-to-end observability flow showing how Aviz processes traffic from Cisco ACI and 5G sources through OPB and Service Nodes for correlation and export.
Workflow showing how Aviz enables application-level visibility by processing tapped data center traffic through OPB and Service Nodes for analysis and export.
Aviz deduplication workflow showing how redundant packets are removed in real time before traffic is forwarded to analytics tools.

Where Does It Shine? Real-World Use Cases Demonstrated

  • Telco Observability: GTP-C and GTP-U correlation at scale, per-user/session KPIs, and 5G SA/NSA analytics.
  • Application Visibility: Dashboards for bandwidth usage, OS/device profiling, geolocation, and threat detection.
  • Data Center Deduplication: 90%+ reduction in traffic from TAPs before it hits monitoring tools—preserving compute and reducing false positives.
  • Natural-Language Queries: Ask questions like “What applications are most used by 5G users?” and receive accurate, traceable answers from NCP.

Is it Built to Scale? Absolutely.

The entire stack is hardware-agnostic and SONiC-ready. OPB supports platforms from Broadcom (Trident, Tomahawk) and NVIDIA (Spectrum 1/2/3). 400G is deployed in production today, and 800G+ support is in active qualification—ensuring that your observability scales with your network.

Conclusion: Visibility Isn’t a Luxury. It’s a Lifeline.

If you’re operating without full network visibility, you’re operating at risk. Aviz Networks offers a scalable, AI-integrated observability stack that empowers NetOps teams to identify issues early, resolve them quickly, and operate with confidence across multivendor environments.

Join our next bootcamp with NVIDIA on June 12 for a deep dive into AI-Powered Networking with Spectrum-X.

Explore how the Open Packet Broker (OPB) delivers real-time traffic filtering, load balancing, and deep observability across disaggregated environments.

Open Packet Broker

Contact Us

Unlocking Deep Network Observability: Lessons from the Field

When it comes to modern networking, what you can’t see can hurt you. At Aviz Networks, we make the invisible visible—before problems escalate. In a recent bootcamp, we explored how deep observability empowers teams to reduce troubleshooting time, achieve compliance, and build more resilient infrastructure. Through a customer case study, a partner perspective, and a […]