What is a FABRIC slice?

A slice is a collection of reserved resources on the FABRIC testbed, including compute nodes, storage, and network links, isolated for specific experimental use.

Is there a way to validate my configuration before deploying?

Yes, the skill utilizes the fablib validation API to check resource availability and configuration logic before submitting the request to the testbed.

Does it handle networking across different sites?

Yes, it includes specific patterns for L3/FABnet configurations that enable seamless communication between nodes located at different physical FABRIC sites.

What happens if a slice creation fails?

The generated code includes robust error handling patterns that automatically attempt to clean up or delete the partial slice to prevent resource leaks.

Can I use this skill to add GPUs to my nodes?

Yes, the skill supports adding components like the RTX6000 and other specialized hardware like FPGAs and SmartNICs during the node creation process.

FABRIC Slice Creator

Name: FABRIC Slice Creator
Author: fabric-testbed

byfabric-testbed

•

クラウドインフラストラクチャ

Automates the creation and configuration of experimental network slices on the FABRIC testbed using Python-based API patterns.

This skill enables Claude to programmatically design, validate, and deploy complex network experiments on the FABRIC testbed. It guides users through specifying node requirements—such as CPU, RAM, and specialized hardware like GPUs or FPGAs—and handles intricate networking setups including Layer 2 segments and Layer 3 FABnet routing. By providing standardized code patterns for error handling, non-blocking submissions, and persistent storage, it streamlines the experimental lifecycle from initial request to a ready-to-use research environment.

主な機能

01Configurable Layer 2 and Layer 3 (FABnet) network topologies

02Automated node provisioning with custom resource specs (CPU, RAM, Disk)

032 GitHub stars

04Support for post-boot configuration and SSH connectivity management

05Support for specialized hardware components like GPUs, SmartNICs, and NVMe

06Built-in validation and robust error handling for slice deployments

ユースケース

01Creating complex Layer 3 network topologies with automated routing across FABRIC sites

02Deploying a multi-site distributed system for network latency research

03Setting up GPU-accelerated environments for distributed AI/ML experiments

主な機能

01Configurable Layer 2 and Layer 3 (FABnet) network topologies

02Automated node provisioning with custom resource specs (CPU, RAM, Disk)

032 GitHub stars

04Support for post-boot configuration and SSH connectivity management

05Support for specialized hardware components like GPUs, SmartNICs, and NVMe

06Built-in validation and robust error handling for slice deployments

ユースケース

01Creating complex Layer 3 network topologies with automated routing across FABRIC sites

02Deploying a multi-site distributed system for network latency research

03Setting up GPU-accelerated environments for distributed AI/ML experiments