AI Notebooks - Reversibility Policy for the Notebook Interface product

Objective

This document describes the reversibility policy for the Notebook Interface product covering the OVHcloud AI Notebooks offer.

This policy aims to implement the general reversibility principles and our compliance with the SWIPO IAAS Code of Conduct for cloud providers.

List of features

Features of the product line fall into three categories:

Core features for which we guarantee migration capacity.
OVHcloud implementations that require adaptation to a new migration environment.
Specific features that cannot be guaranteed for migration as they are related to the OVHcloud environment or involve custom developments.

1. Core features

Function	Description	Available formats	Migration model	Available documentation
Import and export Notebooks	Store notebooks in standard Jupyter (.ipynb) or VS Code Notebook formats including import and export without adaptation	(.ipynb), Python, CSV, TXT	Inbound: Direct download of .ipynb files or conversion via Jupyter/VS Code. Outgoing: Direct export of .ipynb notebooks, reusable on any Jupyter or other compatible environment.	Using data in a notebook via the OVHcloud Control Panel
Support for standard AI frameworks	Pre-installation of environments such as TensorFlow, PyTorch, Scikit-learn, MXNet, Hugging Face, etc.	Notebooks, Python scripts	Inbound: import notebooks/scripts using these frameworks. Outbound: export notebooks/scripts, which can be reused on any platform supporting these frameworks.	Tutorial - Using Tensorboard in a notebook
Manage custom environments	Install libraries and dependencies via pip/conda, specific to each notebook	YAML (environments), requirements.txt	Inbound: Recreating the environment via requirements.txt/environment.yml files. Outbound: Manually export from the package list, adapt, and reinstall on the target environment	Features, Capabilities and Limits
VS Code Integration	Visual Studio Code edition support for Notebooks	Notebooks, Python scripts	Inbound: possible import, but requires adaptation of environment if target does not support VS Code. Outbound: export notebooks/scripts, possibly adjust VS Code configuration on target environment	Features, Capabilities and Limits

2. OVHcloud Implementations

Function	Description	Available formats	Migration model	Available documentation
Compatibility with S3¹ Object Storage	User data are synchronized on an OVHcloud Object Storage S3 once the Notebook has been stopped/deleted. S3 storage can be done with an OVHcloud product or outside of OVHcloud infrastructure.	S3 (objects, files)	Inbound: import data via S3 Object Storage or direct download. Outbound: export files/datasets via S3 Object Storage, which can be reused on any other S3-compatible storage	Features, Capabilities, and Limits Object Storage product reversibility policy
User access management	Access management via the OVHcloud interface, non-standard version	NA	Inbound: Manual configuration of accesses on the interface Outbound: adapt rights management according to the target tool (JupyterHub, VS Code Live Share, etc.).	Use data in a notebook via the OVHcloud Control Panel

3. Specific features

Function	Description	Available formats	Migration model	Available documentation
OVHcloud API, CLI and Python SDK	Notebook management, automation and migration via OVHcloud’s own REST API, CLI AI and Python SDK.	JSON, YAML, Python	Inbound: Tools and management interfaces available by default Outbound: APIs and CLIs will depend on the target environments and may require adaptation.	AI Notebooks - Getting started
Anti-DDoS protection	Anti-DDoS is a set of tools and mechanisms designed to absorb denial of service attacks. It includes traffic analysis, "clean-up" via a specialized network, and mitigation using VAC technology developed by OVHcloud.	N/A	Inbound: The anti-DDoS system is part of our infrastructure and is enabled by default. No action is required Outbound: Order and configure an anti-DDoS solution from the new provider	Anti-DDoS OVHcloud

List of architectures

OVHcloud’s service offer includes managed Jupyter and Visual Studio Code environments, with on-demand CPU/GPU resource allocation. Notebooks are isolated by user and project, with a persistent storage on an S3-compatible Object Storage service. The main AI frameworks (TensorFlow, PyTorch, Scikit-learn, MXNet, Hugging Face) are preinstalled with native integration of data science tools (Matplotlib, Seaborn, Pandas). The architecture supports notebook shutdowns and restarts, automatic data synchronization and multi-user management.

Partner services

The OVHcloud partners concerned are listed in the OVHcloud partners directory under the "Data center expansion and Migration" keywords.

OVHcloud also has a dedicated service: OVHcloud Professional Services.

Cost and fees

Billing is based on 'pay-as-you-go'. There is no specific fees applied on service termination. Stopping or deleting a notebook will stop billing for computing resources. The associated storage (Object Storage) remains billed until data are deleted. Notebooks and data must be exported before deleting and canceling the service.

Data retention after contract termination

After deleting a notebook, OVHcloud automatically synchronizes user data on an S3-compatible Object Storage. Notebooks and files remain accessible on this storage space until the user deletes it manually. Once data has been deleted from this space, it cannot be restored.Therefore, it's imperative to export all the necessary data before decommissioning the service.

¹: S3 is a trademark of Amazon Technologies, Inc. OVHcloud’s service is not sponsored by, endorsed by, or otherwise affiliated with Amazon Technologies, Inc.

¿Le ha resultado útil esta página?