Aviz details network copilot's AI-driven insights for multi-vendor networks

Aviz Network Copilot utilizes open-source Large Language Models to analyze and summarize data from multiple network switches, providing enterprises with real-time insights into network operations and compliance. This approach supports technology leaders in monitoring complex network environments across diverse vendors and network operating systems.

Research overview

Network Copilot is designed as a vendor-neutral platform that integrates data from various network devices using the Aviz ONES system. It collects operational information such as device inventories, network health metrics, traffic patterns, and compliance statuses to underpin its analytic processes.

The platform employs a hybrid method combining fine-tuned Large Language Models with Retrieval Augmented Generation (RAG) techniques, enabling it to generate context-aware, conversational summaries tailored to specific network conditions.

Technical breakdown

The solution supports deployment either as a Standalone (SA) on-premises (on-prem) application utilizing commodity servers with NVIDIA GPUs or as a cloud-native service, offering flexibility to accommodate different enterprise operational models and security postures.

By analyzing key performance indicators like bandwidth usage, latency, and traffic growth, Network Copilot assists network operations teams in forecasting capacity requirements and planning infrastructure upgrades. It produces AI-based recommendations aimed at reducing bottlenecks and improving network efficiency.

Product update

Unlike static monitoring tools that rely on predefined dashboards and alerts, Network Copilot delivers dynamic, conversational insights through an interface comparable to ChatGPT. This enables ad-hoc querying, anomaly detection, and real-time troubleshooting within complex multi-vendor environments.

Continuous traffic and device state monitoring allows the platform to detect unusual behaviors such as spikes in traffic, unauthorized access attempts, and irregular resource consumption. It generates alerts with suggested remediation steps to support network security management.

Upcoming features

Future releases are planned to extend Network Copilot's capabilities to ingest cloud network datasets including APIs, logs, and storage information. These enhancements aim to facilitate monitoring of cloud and hybrid network compliance, application performance, and resource utilization using Generative AI (GenAI) techniques.

The development of Network Copilot incorporates ongoing customer input concerning operational conditions and network performance expectations, resulting in iterative refinements of Artificial Intelligence (AI) model accuracy and relevance for enterprise environments.

This Blog Signals brief provides a fact-based summary of Aviz Network Copilot, highlighting its application of GenAI to deliver conversational, real-time network insights useful for enterprise technology and security teams managing multi-vendor infrastructures.