Service Fabric Documentation

• A Service Fabric is a distributed runtime platform composed of a population of one or more Fibres.
• Each Service Fabric is created from a corresponding fabric profile: this template defines the number and type of Fibres required.
• A Service Fabric may host one or more Systems.

• A Fibre is smart OSGi based agent that runs on the host's Java VM: the host may be a physical machine, a virtual machine, or a container.
• A Fibre may host one or more system.parts.
• A Fibre may act as either:
  • `the container` for one or more OSGi assemblies. 
  • `the proxy / controller` for one or more container images.
  • A combination of each.

The Fibre is responsible for the creation of any required local Linux container (e.g. Docker), and then container image life-cycle management.

• A System is a modular distributed Application. 
• A System is composed of one or more system.parts. 
• Each system.part is composed of one or more colocated system.part.elements.

When deploying a System, the Service Fabric will place the required number of system.part instances upon most appropriate Fibres available. Each instance is dynamically configured by its Fibre host.

Each system.part.element may be the implementation of a CRUD/REST microservice, an Synchronous or Asynchronous RPC microservice, or a consumer / producer of messages or events.

The Service Fabric automatically enforces appropriate visibility / boundaries for each System's microservices:

• Internal microservices are only visible to other entities within the same System.
• External or Public microservices are visible to third parties.