AI Notebooks - Billing and lifecycle

Introduction

AI Notebooks are linked to a Public Cloud project. The whole project is billed at the end of the month, with pay-as-you-go. It means you will only pay for what you consume, based on the compute resources you use (CPUs and GPUs), their running time, and data.

AI Notebooks lifecycle

During its lifetime, the notebook will go through the following statuses:

• STARTING: The notebook is being started and, if any, the remote data is synchronized from the Object Storage. To learn more about data synchronization, please check out the Data - Concept and best practices documentation. Then, the system allocates the necessary compute resources (CPUs/GPUs) for your Notebook. Finally, the base framework you have chosen is pulled for use in the notebook.
• RUNNING: The notebook is running; you can connect to it using its endpoint and benefit from your compute resources and your attached data.
• STOPPING: The notebook is stopping, your compute resources are freed, and your work included in your /workspace folder is automatically saved. If remote data is mounted on your notebook, it is synchronized back to the Object Storage if modified and mounted as RW, not RO.
• STOPPED: The notebook ended normally. You can restart it whenever you want or delete it. It will keep the same endpoint.
• FAILED: The notebook ended in error, e.g., the process in the notebook finished with a non-zero exit code. For more information, refer to this section of our Troubleshooting documentation.
• ERROR: The notebook ended due to a backend error. You may reach out to our support.
• DELETING: The notebook is being removed. When it is deleted, you will no longer see it, and it will no longer exist. Its internal /workspace data will not be deleted automatically after notebook deletion. If you are sure you want to delete it, you can find and delete it in the notebooks_workspace container of your Object Storage, under the notebook ID directory.

Billing principles

AI Notebooks is a pay-per-use solution. You only pay for the resources consumption.

Included in AI Notebooks resources:

• Dedicated CPU/GPU compute resources (based on the selected amount during notebook creation)
• Ephemeral local notebook storage (size depends on the selected compute resources). First 10 GB are free.
• Workspace remote storage (Optional)
• Ingress/Egress network traffic (Optional)

Here is a detailed graph that illustrates every step that is billed or not during the AI Notebook workflow:

Compute resources details

During the notebook creation, you can select compute resources, known as CPUs or GPUs. Their official pricing is available in the OVHcloud Control Panel or on the OVHcloud Public Cloud website.

Compute rates are listed per hour for easier reading, but billing remains per minute, with per-second granularity.

As stated above and shown in the image above, you pay for these resources as long as you consume them. This happens when the image of your notebook is pulled during the STARTING phase, but also during the RUNNING and STOPPING phases, until you reach the STOPPED phase.

Storage details

There are three types of storage within AI Notebooks:

• Remote Object storage
• Workspace storage
• Ephemeral local storage

The pricing of these different storage types is different.

Remote Object storage

Remote data is the type that comes from the OVHcloud Object Storage solution. During notebook creation, you can mount some Object Storage containers into your notebook.

In situations where you are utilizing notebooks with attached remote data, you will be charged separately for the storage of this data. The cost of Object Storage is independent of the pricing for notebooks.

Workspace storage

By default, your notebook will be mounted to a remote Object Storage container on the /workspace location. This will be your default folder when you access your notebook.

You can store your data there.

Your /workspace folder is automatically saved when you stop your notebook. You can also trigger a manual data synchronization if needed.

The first 10GB are free for 30 consecutive days once your notebook is stopped; then you pay at the price of OVHcloud Object Storage.

Ephemeral local storage

Each compute resource (CPU or GPU) comes with local storage, which we can consider ephemeral since this storage space is not saved when you stop or delete your notebook.

The sizing depends on the selected amount of compute resources; check the details on the OVHcloud Public Cloud website.

This concerns locations outside your /workspace, as well as outside any other remote Object Storage containers you may have mounted on your notebook.

This storage is not billed, as it is directly linked to the compute resource(s) you have chosen.

Pricing examples

Example 1: one GPU notebook for 45 minutes, then deleted

We start an AI Notebook with two GPUs, and we keep it running for 45 minutes, then we delete it.

• compute resources: 1 x GPU NVIDIA L4 ($0.91 / hour)
• remote storage: nothing
• duration: 45 minutes

Compute cost:

0.75 (hours) x 1 (GPU) x $0.91 (price / GPU) = $0.6825

Storage cost: none

Total: $0.6825, billed at the end of the month

Example 2: one GPU notebook for 10 hours, but stopped, not deleted

We start an AI Notebook with two GPUs, run it for 10 hours, stop it, and delete it after 10 days. Workspace storage used is 100 GB (the first 10 GB is free).

• compute resources: 2 x GPU NVIDIA L4 ($0.91 / hour)
• remote storage: none
• workspace storage: 90 GB chargeable
• duration: 10 hours running, 10 days (240 hours) stopped

Compute cost:

10 (hours) x 2 (GPU) x $0.91 (price / GPU) = $18.20

Storage cost:

90 GB x 240 (hours) x $0.00001111 (price / hr / GB) = $0.239

Total: $18.44, billed at the end of the month

Example 3: one GPU notebook for 10 hours with 1 TB remote storage

We start an AI Notebook with two GPUs and 1 TB of remote storage. It runs for 10 hours and is then deleted. Workspace usage is 1 TB (990 GB chargeable).

• compute resources: 2 x GPU NVIDIA L4 ($0.91 / hour)
• remote storage: 1,000 GB
• workspace storage: 990 GB chargeable
• duration: 10 hours

Compute cost:

10 (hours) x 2 (GPU) x $0.91 (price / GPU) = $18.20

Remote storage cost:

1,000 GB x 10 (hours) x $0.00001111 (price / hr / GB) = $0.111

Workspace storage cost:

990 GB x 10 (hours) x $0.00001111 (price / hr / GB) = $0.110

Total: $18.42, billed at the end of the month

Example 4: 15 CPU notebooks for 5 hours, then deleted

We start 15 AI Notebooks, each with one vCPU, run them for 5 hours, and then delete them.

• compute resources: 1 x vCPU ($0.04 / hour / CPU)
• remote storage: none
• duration: 5 hours

Compute cost:

5 (hours) x 15 (CPU) x $0.04 (price / CPU) = $3.00

Storage cost: none

Total: $3.00, billed at the end of the month

Images decommissioning

Objective

Products evolve and eventually reach the end of their lifecycle. Our objective is to ensure this transition happens smoothly.

All AI images follow the Python version lifecycle policy. Once a Python version reaches its end of life, the associated AI frameworks will be decommissioned.

What does it mean?

• You will not be able to create new notebooks based on this Python version.
• Your existing notebooks will continue to work until you delete them.
• During this period, the images will still receive security updates.
• Once the underlying OS reaches its end of life, you will still be able to start your notebooks, but the image will no longer be maintained. We strongly encourage you to migrate to a newer image version.

Go further

For more information and tutorials, please see our other AI & Machine Learning support guides or explore the guides for other OVHcloud products and services.

If you need training or technical assistance to implement our solutions, contact your sales representative or click on this link to get a quote and ask our Professional Services experts for a custom analysis of your project.