Skip to main content
Version: v4.6

SMI Requirements overview

Requirements prior to installation#

Before beginning the SMI installation, ensure that your current system has the full set of required system resources which are described in this page.

These requirements are broken down into the following categories:

Essential system resources#

A Digital Twin Platform environment will generally require the following system resources:

  • Core Networking
    • Subnets, routing tables, internet access, etc.
    • Load balancing, WAF, etc. for ingress into Kubernetes
  • Kubernetes v1.29+
    • Network Policy Implementation
    • Custom Metrics Collection Implementation
    • Block storage CSI Provider
    • Shared file system CSI Provider
  • Postgres v14
  • MongoDB v6
  • Redis v7
  • Neo4J v4.4
  • Kafka v3.7
  • ActiveMQ v5.18
  • SMTP
  • DNS

Suggested requirements for Amazon Web Services#

For a detailed list of requirements, refer to the AWS requirements page.

Suggested requirements for Oracle Cloud Infrastructure#

For a detailed list of requirements, refer to the page Oracle requirements page.

Other suggested requirements#

It is also recommended to have the following:

  • Ingress, Naming, and Certificates: Use ingress-nginx, external-dns, and cert-manager to auto-create your cloud native load balancer, proxy your incoming traffic to the Kubernetes cluster, assign it DNS names, and create certificates from Let’s Encrypt or another ACME based certificate provider.
  • External Secrets: Store all your generated secrets in AWS Secrets Manager and then pull those secrets into Kubernetes using External Secrets Operator. This is especially helpful for generating the secrets for external services used by the Digital Twin Platform Helm Chart.
  • Service Mesh: For mTLS connections between services inside Kubernetes. Our installation doesn’t provide native TLS between services and to other Kubernetes installations (such as neo4j). The expectation is some service mesh is in place to encrypt traffic (for example, Istio, Linkerd, AWS App Mesh Controller, etc.)
  • Cluster Autoscaling: Services are configured to horizontally scale when necessary. As such, a cluster autoscaler should be in place (for example, cluster-autoscaler, AWS Karpenter, etc.).