Welcome to Altair AI Cloud!
Our cloud platform should look familiar to you, if you are a previous user of Altair AI Studio / Altair AI Hub.
As before, teams can collaborate when developing data science projects, deploying the results to generate business value. The cloud platform actually doubles down on the idea of a multi-persona, multi-skills environment, but in a new setting where all the hassle of deployments, upgrades, and maintenance goes away.
-
The Altair AI Cloud Migration extension will help you to migrate your projects from Altair AI Studio / Altair AI Hub to our new cloud platform. The extension requires Altair AI Studio 10.3 or later.
-
If you are running a version of AI Studio earlier than 10.3, you can use the standalone migration tool to migrate.
Download the migration tool from here:
Windows: https://releases.rapidminer.com/cloud/migration-tool/v3/ai-cloud-migration-2024.0.3-win64.zip
Linux: https://releases.rapidminer.com/cloud/migration-tool/v3/ai-cloud-migration-2024.0.3-any.zip
MacOS: https://releases.rapidminer.com/cloud/migration-tool/v3/ai-cloud-migration-2024.0.3-osx.dmg
Conceptual differences
We know that many of our existing users will want to move to the cloud platform, so we have tried to maintain compatibility. Nevertheless, due to some architectural improvements, the mapping between old and new is not exact.
One of the changes relates to how data is handled and stored. In Altair AI Studio / Altair AI Hub, data is stored in a local repository, in a remote repository, or in projects. In the cloud, on the other hand, the important concepts are:
- projects and
- data assets.
Projects
Projects are use cases; they are central to development and deployment. Because projects are designed for collaboration, permissions are project-wide.
Within a project, you and your team will create and store anything related to the use case, such as processes, analyses, scripts, and models. In short, a project is the home for your work: where you collaborate, train, and deploy.
Data Assets
Data is treated separately from projects, for several reasons:
- To the extent that data is raw material, it is external to the workflow, not modified by users.
- Nor is data usually deployed together with the project, even if the project reads, transforms, and enriches it.
That's why all data management is now centralized in the Data Assets, with two components:
All the data assets are securely protected via permissions, allowing for sharing and collaboration. They are fully searchable and organized through tags. Data assets can be shared or linked with any number of projects.
When you are working on a project, only the files that are linked to that project are displayed in the Files component.