This image shows the machine learning training and inference workflow within the architecture. The workflow is described in
the surrounding text. The architecture is described below.
This image shows a single compartment and region with a single virtual cloud network (VCN) that spans 2 availability domains,
each with 2 fault domains. The region provides the following services:
- Identity and access management
- Auditing
- Policies
- Logging
- Object storage
The virtual cloud network (VCN) provides the following gateways:
- Dynamic routing gateway (DRG): Provides private connectivity using IPSec VPN or FastConnect for developer remote access.
- Network address translation (NAT) gateway: Enables private resources in a VCN to access hosts on the internet, without exposing
those resources to incoming internet connections.
The VCN includes 5 subnets:
- Bastion public subnet 1: Provides a compute instance that serves as a secure, controlled entry point to the topology from
outside the cloud using a dynamic routing gateway. The bastion host resides in availability domain 1 and fault domain 1.
- API Gateway public subnet 2: This subnet is not used in this workflow.
- Deep resolution inference private subnet 2: Provides deep resolution inference servers in each of the availability domains
with connections to object storage.
- Field delineation inference private subnet 3: Provides a field delineation inference server in each of the availability domains,
a primary database (availability domain 1, fault domain 2), and a secondary database (availability domain 2, fault domain
1).
- Training instances private subnet 1: Provides coupled bare metal and block storage instances in each of the availability and
fault domains.