Skip to content

README.md

Latest commit

 

History

History
177 lines (148 loc) · 7.66 KB

README.md

File metadata and controls

177 lines (148 loc) · 7.66 KB

Legacy Migration

AI-powered legacy system migration analysis that assesses codebases, plans phased migrations, and converts legacy code for financial services institutions.

Overview

The Legacy Migration use case coordinates three specialist agents to provide comprehensive migration assessments for financial services legacy systems. It analyzes codebase structure and dependencies, develops phased migration plans with risk mitigation strategies, and performs code conversion with confidence scoring -- enabling engineering teams and stakeholders to make informed migration decisions.

Business Value

  • Risk-aware planning -- Automated dependency mapping and complexity assessment reduce migration surprises
  • Phased approach -- Migration plans with rollback strategies minimize business disruption
  • Conversion confidence -- Automated code conversion with confidence scoring identifies areas needing manual review
  • Cost estimation -- Effort-day estimates and risk assessments support budget planning and stakeholder alignment
  • Regulatory compliance -- Financial system-specific considerations ensure migration preserves compliance posture

Architecture

graph TB
    Request["Client Request"] --> Runtime["AgentCore Runtime"]
    Runtime --> Orchestrator["Legacy Migration<br/>Orchestrator"]
    Orchestrator --> Agent1["Code Analyzer<br/><small>Codebase structure,<br/>dependency mapping</small>"]
    Orchestrator --> Agent2["Migration Planner<br/><small>Phased migration plans,<br/>risk assessment</small>"]
    Orchestrator --> Agent3["Conversion Agent<br/><small>Code conversion,<br/>pattern modernization</small>"]
    Agent1 --> Bedrock["Amazon Bedrock<br/>(Claude)"]
    Agent2 --> Bedrock
    Agent3 --> Bedrock
    Agent1 --> S3["S3 Sample Data<br/>(Legacy Code)"]
    Agent2 --> S3
    Agent3 --> S3
    Agent1 --> Synthesis["Result Synthesis"]
    Agent2 --> Synthesis
    Agent3 --> Synthesis
    Synthesis --> Response["Response"]
Loading

Directory Structure

use_cases/legacy_migration/
├── README.md
└── src/
    ├── __init__.py                              # Framework router + registry
    ├── strands/
    │   ├── __init__.py
    │   ├── config.py
    │   ├── models.py                            # MigrationRequest / MigrationResponse
    │   ├── orchestrator.py                      # LegacyMigrationOrchestrator
    │   └── agents/
    │       ├── __init__.py
    │       ├── code_analyzer.py
    │       ├── migration_planner.py
    │       └── conversion_agent.py
    └── langchain_langgraph/
        ├── __init__.py
        ├── config.py
        ├── models.py
        ├── orchestrator.py
        └── agents/
            ├── __init__.py
            ├── code_analyzer.py
            ├── migration_planner.py
            └── conversion_agent.py

Agentic Design

The LegacyMigrationOrchestrator extends StrandsOrchestrator and uses a parallel fan-out / synthesize pattern with scope-dependent agent combinations:

  1. Fan-out -- For full scope, all three agents run in parallel via asyncio.gather (async) or run_parallel (sync).
  2. Targeted modes -- code_analysis runs the code analyzer alone; planning pairs code analyzer + migration planner; conversion pairs code analyzer + conversion agent. The code analyzer always runs as a prerequisite for other scopes.
  3. Synthesis -- Agent results are combined using build_structured_synthesis_prompt with a schema covering complexity level, languages detected, estimated effort days, conversion confidence, risks, and recommendations. The orchestrator LLM produces the final migration readiness assessment.

Agents

Code Analyzer

  • Role: Analyzes codebase structure, maps dependencies, and assesses complexity for migration planning
  • Data: Project profile and legacy code from S3 (data_type='profile')
  • Produces: Languages detected, total files and lines, complexity level (low/medium/high/critical), dependency graph, code patterns identified, migration risks
  • Tool: s3_retriever_tool

Migration Planner

  • Role: Develops phased migration plans with risk assessment, dependency ordering, and rollback strategies
  • Data: Project profile from S3
  • Produces: Migration phases with timelines, estimated effort in days, risk assessment, dependency execution order, rollback strategy
  • Tool: s3_retriever_tool

Conversion Agent

  • Role: Performs code conversion from legacy languages/patterns to modern equivalents with confidence scoring
  • Data: Project profile and source code from S3
  • Produces: Files converted count, conversion confidence score (0-1), patterns converted, items needing manual review, target framework
  • Tool: s3_retriever_tool

Data & Tools

Resource Description
s3_retriever_tool Retrieves project profiles, legacy codebases, and documentation from S3
S3 path data/samples/legacy_migration/{project_id}/profile.json

Request / Response

MigrationRequest

Field Type Description
project_id str Project identifier (e.g., PROJ001)
migration_scope MigrationScope full, code_analysis, planning, conversion
additional_context str | None Optional context

MigrationResponse

Field Type Description
project_id str Project identifier
migration_id str Unique migration analysis UUID
timestamp datetime Analysis timestamp
code_analysis CodeAnalysisResult | None Languages, complexity, dependencies, risks
migration_plan MigrationPlanResult | None Phases, effort days, risk assessment, rollback strategy
conversion_output ConversionResult | None Files converted, confidence score, manual review items
summary str Executive summary
raw_analysis dict Raw output from each agent

Example Request:

{
  "project_id": "PROJ001",
  "migration_scope": "full"
}

Example Response:

{
  "project_id": "PROJ001",
  "migration_id": "uuid",
  "timestamp": "2026-03-25T00:00:00Z",
  "code_analysis": {
    "languages_detected": ["COBOL", "JCL", "SQL"],
    "total_files": 1200,
    "total_lines": 450000,
    "complexity_level": "high",
    "dependencies": ["DB2", "CICS", "MQ Series"],
    "risks": ["Tightly coupled batch jobs", "Undocumented business rules"]
  },
  "migration_plan": {
    "estimated_effort_days": 180,
    "risk_assessment": ["Data migration integrity", "Batch job sequencing"],
    "rollback_strategy": "Parallel run with automated comparison for 30 days"
  },
  "conversion_output": {
    "conversion_confidence": 0.72,
    "patterns_converted": ["COBOL PERFORM to Python functions", "COPY to Python imports"],
    "manual_review_needed": ["Complex REDEFINES clauses", "Embedded CICS calls"],
    "target_framework": "Python/FastAPI"
  },
  "summary": "High-complexity COBOL system. 72% automated conversion confidence with 180-day migration estimate."
}

Quick Start

USE_CASE_ID=legacy_migration FRAMEWORK=strands AWS_REGION=us-east-1 \
  ./applications/fsi_foundry/scripts/deploy/full/deploy_agentcore.sh

Sample Data

Project ID Description
PROJ001 Core Banking COBOL System -- 450K LOC, COBOL to Python

Related Documentation