Full Specification

📎 View on Notion →

Overview

Purpose: Describe the multi-stage process of building Unified Models for VM deployment.

Context: MIND Universe VM deployment backend.

Problem: VM deployment requires multi-source, multi-stage model construction.

Architecture

Model States

The Prototype state is built from Source Unit and Placement data sources, tightly coupling the model to infrastructure specifics. User customization transforms the prototype into the Base state, allowing overrides of default values. The Complete state is reached after Universe validation and enrichment with platform-specific default arguments, making the model ready for deployment.

┌───────────────┐      ┌───────────────┐      ┌───────────────┐
│   Prototype   │─────▶│     Base      │─────▶│   Complete    │
│    Model      │      │    Model      │      │    Model      │
└───────┬───────┘      └───────┬───────┘      └───────┬───────┘
        │                      │                      │
        ▼                      ▼                      ▼
┌───────────────┐      ┌───────────────┐      ┌───────────────┐
│ Source Unit + │      │  User Input   │      │  Validation + │
│  Placement    │      │ Customization │      │ Default Values│
└───────────────┘      └───────────────┘      └───────────────┘

Model States

State Source Purpose
Prototype Source Unit + Placement Initial model tied to data sources
Base User customization User overrides prototype values
Complete Universe validation + defaults Ready for deployment

Key Design Decisions

Decision Alternative Rationale
Three-state model Single-state Clear separation of concerns
User customization step Direct deploy User control over final config
Validation before deploy Validate on target Catch errors early

Rollback Mechanism [!]

Note: A rollback mechanism was implemented through DRUID API calls to remove redundant CIRs during deployment. This feature exists in the codebase but cannot be publicly documented due to NDA restrictions.

Trade-offs

Trade-off Impact
Multi-stage process Increased complexity
User customization More flexibility, more validation
Default values Consistency across platforms

← Back to Specifications