COBOL to C# Migration Services

Section 02

Technical Deep Dive

1. Handling CICS BMS Maps

One of the biggest challenges is converting "Green Screens" (BMS Maps) to modern web interfaces.

• The Lazy Way: Use a terminal emulator in the browser. This preserves the code but fails to deliver user value.
• The Modern Way: Extract the screen definitions into a Blazor or React frontend, and expose the COBOL business logic as a REST API.

2. Database Migration: DB2 to SQL Server

Moving from DB2 z/OS to SQL Server or Azure SQL is relatively straightforward due to strong SQL standards support. However, stored procedures and triggers often require manual tuning.

Key Consideration: Ensure your migration tool handles Cursor logic correctly. COBOL cursors behave differently than T-SQL cursors, particularly regarding locking and concurrency.

3. Code Structure: Procedural to Object-Oriented

COBOL is procedural (PERFORM PARAGRAPH). C# is Object-Oriented.

Migration Pattern:

1. Lift and Shift: Convert COBOL programs to C# classes where each paragraph is a method. This creates "C-BOL" (C# written like COBOL).
2. Refactor: Group related methods into Domain Services.
3. Optimize: Replace custom file I/O logic with standard .NET System.IO or Entity Framework.

// Before: C-BOL Style
public void CALCULATE_INTEREST() {
    // procedural logic
}

// After: Domain Driven Design
public class InterestCalculator {
    public decimal Calculate(Account account) {
        // encapsulated logic
    }
}

Section 03

The Hidden Cost: Testing

COBOL systems often have zero automated tests. Before migration, you must:

1. Reverse-engineer test cases from production logs
2. Build regression suite (expect 6-12 months)
3. Run shadow deployments for 3-6 months

Budget 40% of total project cost for testing alone.

Section 04

Red Flag: "We'll refactor later"

Vendors who say "We'll get it working first, then refactor" are lying. Once the migrated system is in production, no one will ever refactor it. Insist on clean code from Day 1.

---

How to Choose a COBOL to C# Migration Partner

If you need a 100% automated conversion: TSRI. Their model-based engine guarantees 100% functional equivalence with zero manual code intervention.

If you want to reduce MIPS costs immediately: CloudFrame. They allow you to cross-compile COBOL to Java/C# and run it on zIIP processors (cheaper) while keeping the source as COBOL.

If you are moving to Azure: Astadia. They have a specialized "Mainframe to Azure" factory model and deep partnership with Microsoft.

If you have a complex legacy database (IDMS, Datacom): Advanced Modernization. They specialize in the hard stuff that others won't touch.

Red flags:

• Vendors who propose a "Big Bang" cutover for a system >1M LOC
• No automated testing strategy (Record/Replay)
• Lack of experience with your specific COBOL dialect (Micro Focus vs IBM Enterprise)
• "Black Box" conversion tools that produce unreadable C# code

---

Section 05

When to Hire COBOL to C# Migration Services

1. Azure Strategy Mandate

Your CIO has declared "All in on Azure." The mainframe is the last outlier in an otherwise .NET shop.

Trigger: Data center closure deadline; Azure consumption commitment.

2. Performance Bottlenecks

Batch windows are bleeding into the online day. You can't scale the mainframe vertically without spending millions on hardware upgrades.

Trigger: SLA breaches; inability to process nightly batch within 4 hours.

3. Integration Paralysis

Your new C# web apps need data from the mainframe, but building SOAP/REST wrappers around CICS transactions is slow and fragile.

Trigger: "It takes 3 weeks to add one field to the API."

4. Compliance & Audit

Regulators demand real-time reporting and granular data access that VSAM files cannot provide efficiently.

Trigger: Failed audit; regulatory fine risk.

5. Cost Reduction

Mainframe MLC (Monthly License Charges) are consuming 60% of the IT budget.

Trigger: CFO demands 30% OpEx reduction.

---

Section 06

Total Cost of Ownership: Mainframe vs Azure

Hidden Costs NOT Included:

• Parallel Run: Running Mainframe + Azure simultaneously for 3-6 months (double cost).
• Data Egress: Moving petabytes of historical data out of the mainframe environment.
• Third-Party Software: Replacing SyncSort, CA-7, and other mainframe utilities with Azure equivalents.

Break-Even Analysis:

• Median Investment: $3M
• Annual Savings: $1.2M (MLC + Hardware)
• Break-Even: 2.5 years

---

Section 07

COBOL to C# Migration Roadmap

Phase 1: Assessment & Proof of Concept (Months 1-3)

Activities:

• Run code analysis to identify "Dead Code" and complexity
• Select a representative slice (e.g., 50 programs) for a Pilot
• Prove the conversion tool produces maintainable C#
• Validate performance on Azure

Deliverables:

• Pilot Application Running on Azure
• Conversion Factory Setup
• Detailed Project Plan

---

Phase 2: Database & Data Migration (Months 4-8)

Activities:

• Convert DB2 DDL to Azure SQL / SQL Server
• Migrate VSAM files to Blob Storage or SQL Tables
• Build CDC (Change Data Capture) pipelines for sync

Risks:

• EBCDIC to ASCII conversion issues
• Stored Procedure logic mismatch

Deliverables:

• Target Database Live
• Data Sync Active

---

Phase 3: Application Conversion (Months 9-20)

Activities:

• Mass conversion of COBOL to C#
• Rewrite CICS BMS maps to ASP.NET Core / Blazor
• Replace JCL with Azure Logic Apps or Control-M
• Unit Testing

Deliverables:

• C# Codebase
• Modern UI
• Batch Job Definitions

---

Phase 4: Testing & Cutover (Months 21-26)

Activities:

• Record & Replay Testing: Capture mainframe inputs/outputs and verify C# matches exactly.
• Performance Tuning (Azure SQL Hyperscale)
• Parallel Run
• Go-Live

Deliverables:

• Production System
• Decommissioned Mainframe

---

Section 08

Architecture Transformation

---

Section 09

Post-Migration: Best Practices

Months 1-3: Hypercare

• Performance: Monitor Azure SQL "DTU" consumption. Mainframe I/O is incredibly fast; SQL Server might need tuning (In-Memory OLTP).
• Error Handling: Mainframe errors are cryptic (S0C7). Ensure C# exceptions are logged to Application Insights with full context.

Months 4-6: Refactoring

• Remove "C-BOL": Start refactoring the generated C# code. Replace procedural patterns with proper Object-Oriented design.
• DevOps: Implement a full CI/CD pipeline (GitHub Actions / Azure DevOps) to enable daily releases.

---

Section 10

Expanded FAQs

Will the converted C# code be readable?

Answer: It depends on the tool. TSRI and Astadia produce "maintainable" code, but it will still look like "C-BOL" (C# with COBOL structure) initially. You need a post-migration refactoring strategy to turn it into idiomatic C#.

How do we handle JCL (Job Control Language)?

Answer: JCL is the glue that holds batch jobs together. We typically convert JCL to PowerShell scripts, Python, or use a modern scheduler like Control-M or Azure Logic Apps to orchestrate the C# executables.

Can we keep using VSAM files?

Answer: Technically yes, with emulation drivers, but don't. Migrate VSAM data to Azure SQL or Cosmos DB. Keeping VSAM on the cloud negates the benefits of data accessibility and analytics.

What about "Decimal Precision"?

Answer: COBOL's COMP-3 is exact. C#'s decimal type is also exact (128-bit). Do not use double or float for financial data, or you will fail compliance audits due to rounding errors.

Is "Big Bang" the only way?

Answer: No. For massive systems, we recommend a Co-existence strategy. Migrate the database first, then migrate applications module by module (e.g., "Claims" first, then "Billing"), keeping the rest on the mainframe until finished.

---

Section 11

Risk Factors

Decimal Precision vs. Performance Trade-off

COBOL's COMP-3 (packed decimal) has exact precision for financial calculations. C#'s decimal type is accurate but 10-20x slower than float. Vendors who use double/float will introduce rounding errors in currency calculations - a regulatory compliance disaster.

User Interface Logic (BMS to ASP.NET)

CICS BMS maps are character-based. Modern ASP.NET Core or Blazor apps are event-driven and stateless. The logic governing screen flow often needs a complete rewrite, not just a migration.

---

Section 12

Feasibility Analysis

Ideal candidates

• Organization is already invested in Microsoft/Azure ecosystem
• Desire to leverage .NET developer talent pool
• Need to integrate with modern C# web apps

Break-even cost: $450k Talent risk: High. C# devs are common, but those who understand mainframe patterns are rare.