// Interactive Architecture Diagram

The Sovereign Forge pipeline.

Scroll horizontally on smaller screens. Solid arrows show data flow between phases. Dashed arrows show infrastructure dependencies.

%%{init: {'theme': 'dark', 'themeVariables': {'primaryColor': '#162236', 'primaryTextColor': '#EFF3F8', 'primaryBorderColor': '#1E2E45', 'lineColor': '#0078D4', 'secondaryColor': '#0F1B2D', 'tertiaryColor': '#0A1220', 'edgeLabelBackground': '#0A1220', 'clusterBkg': '#0F1B2D', 'clusterBorder': '#1E2E45'}}}%%

flowchart TB
    subgraph INPUT["๐Ÿ”น INPUT โ€” Legacy COTS Application"]
        direction LR
        INSTALLER["๐Ÿ“ฆ Installer Package\nMSI / NSIS / Inno Setup / InstallShield"]
        BINARY["โš™๏ธ Application Binaries\nPE / MZ / NE / DLLs"]
        DATABASE["๐Ÿ—„๏ธ Database\nSQL Server / Access / FoxPro"]
        USERS["๐Ÿ‘ฅ Subject Matter Experts\nAdmins / End Users"]
    end

    subgraph PHASE1["๐Ÿ”ต PHASE 1 โ€” Discovery"]
        direction LR
        GHIDRA["๐Ÿ”ฌ Ghidra Server\nHeadless binary analysis\nFunction decompilation\nCall graph generation"]
        EXTRACT["๐Ÿ“‹ Installer Extractor\n7z / innounp / unshield\nFile inventory"]
        SCHEMA["๐Ÿ—ƒ๏ธ Schema Recovery\nTable relationships\nStored procedures\nData type mapping"]
        REGDIFF["๐Ÿ“ Registry & FS Diff\nBefore/after snapshots\nCOM registrations"]
        NETCAP["๐ŸŒ Network Capture\ntcpdump / Wireshark\nProtocol analysis"]
    end

    subgraph PHASE2["๐Ÿ”ต PHASE 2 โ€” Knowledge Capture"]
        direction LR
        INTERVIEW["๐ŸŽค Interview Engine\nGuided questionnaires\nRole-specific workflows"]
        SCREEN["๐Ÿ“น Screen Recorder\nSession capture\nUI element annotation"]
        WORKFLOW["๐Ÿ—บ๏ธ Workflow Mapper\nProcess flow diagrams\nRole-based patterns"]
        RULES["๐Ÿ“ Business Rule Extractor\nLLM-driven extraction\nValidation logic"]
        WHISPER["๐ŸŽง Transcription\nWhisper on GPU\nLocal inference"]
    end

    subgraph PHASE3["๐ŸŸก PHASE 3 โ€” Analysis & Decision"]
        direction LR
        DASHBOARD["๐Ÿ“Š BI Dashboard\nComplexity scoring\nRisk heat maps\nWorkflow coverage"]
        RECOMMEND["๐ŸŽฏ Target Recommender\nCost/effort/risk modeling\n4 modernization paths"]
        GAPANALYSIS["๐Ÿ” Gap Analyzer\nUndiscovered workflows\nMissing business rules"]
        COMPLIANCE["๐Ÿ›ก๏ธ Compliance Mapper\nCMMC / HIPAA\nPCI-DSS / NIST"]
    end

    subgraph TARGETS["๐Ÿ“Œ MODERNIZATION TARGETS"]
        direction LR
        SAAS["โ˜๏ธ SaaS\nReplacement"]
        WEBAPP["๐ŸŒ Internal\nWeb App"]
        FATCLIENT["๐Ÿ–ฅ๏ธ Fat Client\n1:1 Rebuild"]
        HYBRID["โšก Hybrid\nWeb + Containers"]
    end

    subgraph PHASE4["๐ŸŸข PHASE 4 โ€” Generation & Validation"]
        direction LR
        CODEGEN["๐Ÿ’ป Code Generator\nSemantic Kernel\nAutoGen Agents"]
        OPENAPI["๐Ÿ“‹ OpenAPI Generator\nAPI surface definition\nData contracts"]
        TESTS["๐Ÿงช Test Suite Builder\nIntegration tests\nRegression tests"]
        CONTAINER["๐Ÿ“ฆ Container Packager\nDocker / OCI\nHelm charts\nAKS manifests"]
        VALIDATE["โœ… Regression Validator\nSide-by-side comparison\nBehavioral verification"]
    end

    subgraph OUTPUT["๐Ÿ”น OUTPUT โ€” Modernized Application"]
        direction LR
        SPECS["๐Ÿ“‹ OpenAPI Specs"]
        IMAGES["๐Ÿ“ฆ Container Images"]
        SUITES["๐Ÿงช Test Suites"]
        GAPREPORT["๐Ÿ“Š Gap Reports"]
    end

    subgraph INFRA["โฌ› INFRASTRUCTURE โ€” Azure Local"]
        direction LR
        AZLOCAL["๐Ÿ—๏ธ Azure Local\nHyperconverged\nOn-premises"]
        HYPERV["๐Ÿ’ฟ Hyper-V Sandbox\nIsolated VMs\nCheckpoint/Replay"]
        ARC["๐Ÿ”— Azure Arc\nControl plane\nPolicy enforcement"]
        AKS["โ˜ธ๏ธ AKS Hybrid\nKubernetes\nContainer orchestration"]
        AIEDGE["๐Ÿง  Azure AI on Edge\nGPU inference\nVision / LLM / Whisper"]
        GPU["๐ŸŽฎ GPU Node\nNvidia A2 / T4\nLocal AI workloads"]
    end

    INPUT --> PHASE1
    INSTALLER --> EXTRACT
    BINARY --> GHIDRA
    DATABASE --> SCHEMA
    USERS --> PHASE2

    PHASE1 -->|"System manifest\nBinary inventory\nSchema maps"| PHASE2
    PHASE2 -->|"Workflow specs\nBusiness rules\nTranscripts"| PHASE3
    PHASE3 --> TARGETS
    TARGETS -->|"Target selection\nFunctional specs\nGap report"| PHASE4
    PHASE4 --> OUTPUT

    INFRA -.->|"Compute & AI"| PHASE1
    INFRA -.->|"GPU inference"| PHASE2
    INFRA -.->|"Processing"| PHASE3
    INFRA -.->|"Build & Deploy"| PHASE4

    HYPERV -.->|"Legacy app\nexecution"| PHASE1
    AIEDGE -.->|"Vision / Whisper\nLLM inference"| PHASE2
    AKS -.->|"Deploys"| OUTPUT

    style INPUT fill:#162236,stroke:#0078D4,stroke-width:2px,color:#EFF3F8
    style PHASE1 fill:#0F1B2D,stroke:#0078D4,stroke-width:2px,color:#EFF3F8
    style PHASE2 fill:#0F1B2D,stroke:#00A4EF,stroke-width:2px,color:#EFF3F8
    style PHASE3 fill:#0F1B2D,stroke:#FFB900,stroke-width:2px,color:#EFF3F8
    style TARGETS fill:#162236,stroke:#C8D1DC,stroke-width:1px,color:#EFF3F8
    style PHASE4 fill:#0F1B2D,stroke:#7FBA00,stroke-width:2px,color:#EFF3F8
    style OUTPUT fill:#162236,stroke:#7FBA00,stroke-width:2px,color:#EFF3F8
    style INFRA fill:#0A1220,stroke:#005A9E,stroke-width:2px,color:#EFF3F8
            
Phase 1 & 2 โ€” Discovery & Knowledge Capture
Phase 3 โ€” Analysis & Decision
Phase 4 โ€” Generation & Output
Infrastructure โ€” Azure Local
Infrastructure dependency (dashed)
// Data Flow Summary

What flows between each phase.

01โ†’02

Discovery โ†’ Knowledge Capture

System manifest, binary inventory, database schema maps, dependency graph, installation artifact catalogue, network protocol captures.

02โ†’03

Knowledge Capture โ†’ Analysis

Workflow specifications, business rule documentation, interview transcripts, screen recording annotations, role-based process maps.

03โ†’04

Analysis โ†’ Generation

Modernization target selection, functional specifications, compliance gap report, risk assessment, cost/effort model.

04โ†’๐Ÿš€

Generation โ†’ Deployment

OpenAPI specifications, Docker/OCI container images, Helm charts, AKS deployment manifests, test suites, comprehensive gap report.

// Zero Egress Architecture

Every byte stays on your network.

Compute

Azure Local hyperconverged infrastructure provides all processing power. Hyper-V sandbox VMs isolate legacy application execution with checkpoint/replay.

AI Inference

Nvidia A2/T4 GPUs run all AI workloads locally โ€” Azure AI Vision for screen annotation, Whisper for transcription, quantized LLMs for code generation.

Management

Azure Arc provides the control plane for policy, monitoring, and updates. Management telemetry is the only outbound traffic โ€” application data never leaves.

Your legacy. Your code. Your infrastructure.

โ† Back to Platform Overview