- sara.next
- Technology
- Deployment models
Technology · Architecture
Deployment models
A single container image contains the backend, the frontend and the runtime. The same version runs in every environment — only the configuration differs.
- Docker
- Kubernetes
- on-premises
- cloud
- from working documentation — to be approved
Topology
What runs where
01 · Users
- BrowserHTTPS
- Identity providerSAML 2.0 sign-in
02 · sara.next container
- FrontendAngular SPA
- BackendSpring Boot · Java 17
- RuntimeJava in the image
one image · external configuration
03 · Services
- DatabaseSQL Server or PostgreSQL
- SMTP servernotifications
- Target systemsREST · Graph · OData · SQL
access from the backend only
Expert view
One image, many environments — what it means in practice
For infrastructure teams three questions matter: where will it run, how is it updated and what do we have to maintain. The container model simplifies the answer to all three.
The application is delivered as a single image containing the backend, the frontend and the runtime. There is no separate application server to install and no set of dependencies to keep track of. An update means starting a new image version with the same configuration.
What stays on the organisation's side
Two dependencies: a relational database and an identity provider. This is a deliberate architectural choice. A database is a resource organisations already know how to maintain, secure and back up according to their own standards. The identity provider, in turn, is where the organisation enforces its own sign-in policy — including multi-factor authentication.
For bodies that cannot or do not want to process entitlement data outside their own infrastructure, this means that an on-premises deployment is not a cut-down version. It is the same application.
Where
No lock-in to an infrastructure provider
You need the ability to run containers and access to a database and an identity provider.
Cloud
A cloud platform that supports containers.
Kubernetes
Deployment in a cluster, with configuration in the deployment definition.
Virtual machines
A server with Docker, e.g. using docker-compose.
On-premises
The organisation's own infrastructure.
Getting started
From image to first sign-in
Database
Preparing a relational database and an account for the application.
Identity provider
A realm, a SAML client for the application and a client for user synchronisation.
Configuration
Environment variables: database, SSO, user synchronisation, mail.
Container start
The database schema and the administrator account are created at initialisation.
First sign-in
The administrator signs in via SSO and grants permissions to further people.
Application onboarding
Dictionaries, applications, roles and import of existing access.
Configuration
Configuration outside the image
Parameters are supplied when the container starts, so the image does not need rebuilding between environments.
System requirementsWays to supply configuration
- Container environment variables
- Configuration files attached to the container
- Parameters in the deployment definition: docker run, docker-compose, Kubernetes, cloud platform
- Sensitive data — in a secrets management mechanism, not in the image
Environments
Test and production with the same process configuration
Separate environments
Development, test and production with their own parameters.
Moving configuration
Roles, rules, reports and notifications are exported and imported between environments.
Image versions
Images are tagged, which lets you specify the exact version in a deployment.
See sara.next on your own scenarios
During the presentation we will walk through an access request, approval, automatic execution of the change and a review — using examples from your organisation.