Edge 372 architecture and installation pages

docs/architecture.md

@@ -0,0 +1,25 @@
+# Architecture
+
+## Summary
+
+A detailed view of decentralized control plane outlining major functions is provided below:
+
+![Components](/components.png)
+
+The DCP performs three functions:
+
+- Application management - responsible for local and global application management making sure that application is running in the current zone
+- Mesh management - responsible for peer state synchronisation across clusters making sure that application state is synchronised
+- Optimal decision making - responsible for optimal execution of workload and decision making
+
+The optimization/decision part is extensible and multiple factors can be taken into account. At the moment the placement controller considers only the classical request/limits for applications pods.
+
+## Details
+
+Detailed documentation regarding each component is available via Deep Wiki:
+
+AnyApplication Controller [DeepWiki](https://deepwiki.com/HIRO-MicroDataCenters-BV/anyapplication-controller)
+
+Placement Controller [DeepWiki](https://deepwiki.com/HIRO-MicroDataCenters-BV/placement-controller)
+
+Mesh Controller [DeepWiki](https://deepwiki.com/HIRO-MicroDataCenters-BV/mesh-controller)

docs/index.md

@@ -1,6 +1,6 @@
 # Decentralised Control Plane for Containerised Applications
 
-# Overview
+## Overview
 
 Decentralized Control Plane for kubernetes applications enables application deployment into a mesh of disjoint kubernetes clusters with intelligent orchestration that considers various optimization criteria - such as application energy consumption, dynamic energy prices, data security zones etc.
 
@@ -12,7 +12,7 @@ Each DCP zone is an autonomous peer that replicates entire state and its updates
 
 The intelligent orchestration enables zone-local decisions taking into account global state in all zones. The orchestration system is extensible with a variety of decision making algorithms.
 
-# Features
+## Features
 
 - Application orchestration in a mesh of disjoint kubernetes clusters
 - Optimal application placement decision and continuous decision review

docs/usage/installation.md

@@ -0,0 +1,44 @@
+# Installation
+
+## **Dependencies**
+
+Requirements:
+
+- Kubernetes cluster ≥ 1.32
+- Cilium or Calico networking with BGP peering (or exposing services externally via LoadBalancers)
+
+## Install Guide
+
+1. Generate multi cluster configuration using script [here](https://github.com/HIRO-MicroDataCenters-BV/anyapplication-controller/tree/main/config/generator)
+
+2. Execute install scripts for all clusters
+
+    TBD
+
+## Application Deployment
+
+The application is deployed via Kubernetes Resource where the application is referenced by its helm chart.
+
+```json
+---
+apiVersion: dcp.hiro.io/v1
+kind: AnyApplication
+metadata:
+  name: my-app
+  namespace: default
+spec:
+  source: 
+    helm:
+      repository: https://helm.app.com/stable
+      chart: my-app
+      version: 1.0.0
+      namespace: default
+      values: |-
+          # values overrides
+  zones: 1
+  placementStrategy: 
+    strategy: Global
+  recoverStrategy:
+    tolerance: 1
+    maxRetries: 3
+```