Architecture
Predibase is a cloud-native, SaaS platform that operates out of a controlplane for business logic and a dataplane for operations related to customer-sensitive data.
The controlplane follows a multi-tenant architecture and is managed by Predibase. The dataplane is a secure Predibase environment which can be deployed to either AWS or Azure cloud. The dataplane is a fully managed, secure Predibase environment that is deployed within your AWS or Azure cloud account. To ensure that all data and API requests remain within your cloud account and never traverse the public internet, a private link is established between your applications and the Predibase dataplane running in your cloud.
Deployment Options
Predibase Cloud
Predibase’s fully managed deployment allows you to start training models instantly on the Predibase platform. Predibase will operate the end-to-end infrastructure stack, including both the controlplane and dataplane. This allows us to easily manage and monitor the infrastructure, while also providing improved operational support for customers.
During the SaaS free trial, you will receive access to the Predibase platform (with limitations) at absolutely no cost.
VPC
The VPC deployment option gives you control over your data while benefiting from the agility of a SaaS solution. This is achieved by deploying the dataplane into your own Virtual Private Cloud (VPC) so that data never leaves your environment. We suggest creating a new cloud account to host the Predibase dataplane to simplify billing and security.
Deploying the dataplane is simple. First, we will ask you to apply a CloudFormation (or Azure Resource Manager) template to your target account. Once that is done, Predibase takes over and fully manages deploying both the infrastructure and applications to run our dataplane. All future upgrades and updates are automatically handled by Predibase.
Direct Ingress
As a best practice, we recommend creating a new cloud account to host the Predibase dataplane. If you require private internet connectivity between your Predibase dataplane and internal VPCs, then we can help you configure a private link between the two accounts. We support the following options:
- AWS PrivateLink
- Azure Private Link
- GCP DirectConnect
Cloud Providers
Users can deploy Predibase into Amazon Web Services (AWS) or Microsoft Azure. While both cloud providers have a number of regions, the curated list below are regions that have our recommended GPUs available.
AWS
| Region | Name | Availability |
|---|---|---|
| us-east-1 | N. Virginia | Available |
| us-east-2 | Ohio | Available upon Request |
| us-west-1 | N. California | Not Recommended |
| us-west-2 | Oregon | Available |
| ap-northeast-1 | Tokyo | Available upon Request |
| eu-central-1 | Frankfurt | Available upon Request |
| eu-south-2 | Spain | Available upon Request |
Azure
| Region | Name | Availability |
|---|---|---|
| us-east | East US | Available upon Request |
| us-south-central | South Central US | Available upon Request |
| us-west-2 | West US 2 | Available |
| europe-west | West Europe | Available upon Request |
| australia-east | Australia East | Available upon Request |
If you’d like to deploy into another cloud (ex. Google Cloud Platform) or another region, please let us know at support@predibase.com.