AIDC Fabric
Network Controller for AI Fabric and Cloud Data Center
Network Controller for AI Fabric and Cloud Data Center
The AI Fabric Controller is a high-performance network management platform specifically designed for AI data centers. It supports full-scenario deployment
across backend, frontend, storage, and DC-converged networks.
AIDC Fabric
Asteria AIDC Controller
Asteria AIDC Controller
- Smart Pre-design
- ZTP
- Multi-Dimensional Monitoring
- One-Click Updates
- Intelligent Alarm
- Periodic Inspection
Centralized management for your entire campus infrastructure, from APs
and switches to routers and integrated PON networks.
Asteria AI Fabric Controller
Function Highlights
Asteria AI Fabric Controller
Function Highlights
ZTP
- Basic network setup (IP, BGP, etc.)
- Batch import via CSV (RoCE, routing, ARS)
- Enable advanced AIDC features
Multi-Dimensional Monitoring
- Network Health Dashboard
- In-depth Telemetry
- Traffic and Packet Loss Insights
One-Click Updates
- Unified Firmware Management
- Incremental Patch Updates
Intelligent Alarm
- Multi-Level Threshold Alerts
- Real-Time Remote Notifications
- Fault Lifecycle Management
Periodic Inspection
- One-Click Service Inspection
- Scheduled Inspections
- Health Assessment Reports
Hands-on Demo
Efficient configuration and intelligent network monitoring.
Watch this step-by-step tutorial to configure our current features.
Stay tuned for our upcoming solution: deploy with EasyRoCE, monitor via INT/Grafana, and utilize EID for intelligent fault resolution and automated operational refinement.
AI Fabric Controller Supported Features
|
Frequently Asked Questions
Frequently Asked Questions
What is the primary function of the AIDC Controller and its typical use cases?
The AIDC Controller is a high-performance management platform designed for AI Data Centers. It automates deployment across four main scenarios:
-
Backend Network: High-bandwidth, low-latency Spine-Leaf architectures.
-
Frontend Network: Service isolation and reliability via EVPN-MCLAG.
-
Storage Network: Lossless data transfer using RoCE and MCLAG.
-
DC Convergence Network: Standard Spine-Leaf networking supporting both EVPN-MCLAG and EVPN-Multi-homing
What are the hardware requirements for the controller?
We recommend deploying on an x86 server running Ubuntu 18.04 LTS (or higher) with Docker 20+.
Asterfusion CX-N series switches natively support the deployment of this controller.
Is the deployment complicated? Does it support containers?
Not at all. It supports one-click installation and is fully containerized. Just upload the
.bin file to your server, run the installation command to set the Controller IP, and use docker-compose up -d to launch the full suite of microservices.
I have hundreds of switches. How do I add them to the controller quickly?
You can follow a simple two-step process:
-
Bulk Import: Upload an Excel/CSV template containing device MAC addresses, models, and Loopback IPs.
-
Auto-Onboarding: Automatically plan the topology for a selected scenario. Once all devices in the planned topology are online and connected, the controller will discover and identify them automatically.
Can I get immediate notifications if the network fails?
Yes. The controller features robust Alarm Management covering dozens of metrics, including:
-
Interface Up/Down status and Optical Power anomalies.
-
IP/MAC flapping.
-
Hardware issues (Fan/Power Supply failures).
Can the controller automate routine maintenance and inspections?
Yes. The “Automated Inspection” feature removes the need for manual checks:
-
Scheduled Inspections: Set tasks to run daily, weekly, or monthly. The system runs quietly in the background.
-
One-Click Health Checks: Instantly perform a deep “physical exam” on specific network areas.
-
Inspection Reports: Detailed records are generated automatically and can be exported as PDFs for your maintenance records.