Skip to content

Pre-work Overview

Complete Before Workshop

The pre-work must be completed before attending the workshop. Deployment can take 30-60 minutes depending on your environment and network speed.

What You'll Do

In this pre-work section, you will:

  1. ✅ Verify you have all prerequisites installed
  2. ✅ Clone the workshop repository (includes all lab materials)
  3. ✅ Choose a deployment option (local or cluster)
  4. ✅ Deploy Geospatial Studio in your environment
  5. ✅ Verify the deployment is working correctly

Time Required

  • Prerequisites check: 10 minutes
  • Local deployment: 30-45 minutes
  • Cluster deployment: 45-60 minutes
  • Verification: 10 minutes

Total: 1-1.5 hours

Deployment Options

You have two deployment options:

Deploy Geospatial Studio on your local machine using Lima VM (macOS/Linux).

Pros: - No cloud resources required - Good for learning and testing - Easier to troubleshoot - Simple, pre-configured setup

Cons: - Limited performance (no GPU acceleration) - Requires significant local resources - Not suitable for production workloads - Fixed in-cluster services only (PostgreSQL, MinIO, Keycloak)

Service Configuration: - All services automatically deployed within Lima VM - No option to use external cloud services - Simplified setup for learning and testing

Best for: Workshop participants, developers, testing

Local Deployment Guide →

Option 2: Cluster Deployment

Deploy on Red Hat OpenShift or Kubernetes cluster with GPU support.

Pros: - Full performance with GPU acceleration - Production-ready - Scalable - Flexible service configuration (in-cluster or external cloud services)

Cons: - Requires cluster access - More complex setup - May incur cloud costs

Service Configuration: - Choose between in-cluster services (PostgreSQL, MinIO, Keycloak) - OR external cloud-managed services: - IBM Cloud Databases, AWS RDS, Azure PostgreSQL, GCP Cloud SQL - IBM COS, AWS S3, Azure Blob Storage, GCP Cloud Storage - IBM Security Verify, external Keycloak, Okta, Azure AD

Best for: Production deployments, GPU-intensive workloads, enterprise use

Cluster Deployment Guide →

What Gets Deployed

The Geospatial Studio deployment includes:

Component Purpose Configuration Options
Studio Gateway API Main API for all backend services Always deployed
Studio UI Web-based user interface Always deployed
PostgreSQL Database for metadata storage In-cluster (local) OR cloud-managed (cluster)
MinIO S3-compatible object storage In-cluster (local) OR external S3 (cluster)
Keycloak OAuth2 authentication In-cluster (local) OR external OAuth (cluster)
MLflow Experiment tracking for model training Always deployed
GeoServer Geospatial data visualization Always deployed
Redis Caching and message queuing Always deployed

Service Configuration Flexibility

Local deployment uses fixed in-cluster services with no configuration options.

Cluster deployment offers flexible configuration - you can choose between in-cluster services or external cloud-managed services (IBM Cloud, AWS, Azure, GCP) for PostgreSQL, object storage, and authentication.

Architecture Details

For a detailed architecture overview with diagrams and component descriptions, see the Architecture Overview section.

Next Steps

  1. Check Prerequisites →
  2. Choose your deployment option:
  3. Verify Installation →

Need Help?

If you encounter issues during deployment:

Next: Prerequisites →