Saturn Cloud is a greate choice for data science teams who want to maximize flexibility of their environment. Integrated parallel processing with Dask differentiates it from the competitors. With open source tools teams can design a workflows that fit best to their specific needs.
Saturn Cloud hosting and pricing
The hosted product is available for individuals as well as organizations. The organization version gives better tools for resource management. Enterprise version is deployable from AWS marketplace.
Saturn cloud has three pricing tiers.
First is called Hosted and it is for individual data scientists. You get 30 free compute hours per month on a powerful machine with tools like JupyterLab, RStudio, and Dask. Additional hourse follow the pay-as-you-go pricing.
The Hosted Organization plan has same development tools than the individual plan. But additionally the team gets an admin panel to manage users, access and compute resources.
Satrun Cloud takes 30-60% surplus on top of raw AWS compute prices. This level of service fee is typical.
In the Enterprise plan you can install Saturn Cloud directly from AWS marketplace. You get technical support, single sign-on and installation to custom cloud VPC.
Read more here about Saturn Cloud pricing.
Dask is the competitive edge for Saturn Cloud
Distributed computation is possible with integrated Dask. Managed Dask is the major distinctive factor from competitors
Saturn Cloud also co-operates with Bodo to manipulate data in one server but multiple cores.
Saturn Cloud has many similarities with Datalore. Even though, Saturn Cloud might be a step behind in the developer experience.
Developer tools in Saturn Cloud
Data scientists can connect to the compute clusters from local IDEs such as VS Code via SSH.
Git functionality is built-in.
You can publish an API or a dashborad and schedule the deployments.
It is possile to run deployments and jobs from a custom Docker image.
Programming languages supported by Saturn Cloud
While supporting Python and R language, Saturn Cloud provides also Julia out of the box. With custom containers you can use any programming language or framework you like!
Summary of Saturn Cloud
Saturn Cloud claims to be extremely flexible platform. It redeems the user from fixed workflows. It can be seen more as an data science platform rather than MLOps one.
The storage and computation are disconnected so that data scientist can enjoy productive work days without extra costs.
Saturn Cloud has three types of resources:
- Workspace for exploratory data science work
- Deployments can be done to serve ML models through an API
- Jobs to run batch processing on schedule