Add mainframe-exit migration plan for IPMS workload

[devin-ai-integration]

Summary

Adds MIGRATION_PLAN.md — a planning/architecture report recommending how to migrate the Investment Portfolio Management System off IBM i / z/OS. No code changes; single new markdown doc.

What it answers (the core sequencing question — "service-level vs data conversion first?"):

• Rejects both wholesale "data-first big-bang" and "service/logic-first with a legacy data bridge" as banking-unsafe (data gravity, cross-platform ACID/2PC, blast radius, rollback).
• Recommends a domain-by-domain Strangler Fig with a precise intra-domain order: data foundation first (ACL + reconciliation harness, then CDC/dual-write while the mainframe stays system-of-record), service migration second, and write-authority cutover last (reads cut over before writes, per domain).

Grounded in the actual repo inventory (counts verified from the tree/grep): 11 batch, 8 online/CICS, 8 portfolio, 6 common, 3 utility, 2 test programs; 20 copybooks; 15 JCL; DB2 (EXEC SQL×14) + CICS (EXEC CICS×8) + COMP-3×7 + VSAM ASSIGN×25. Flags real hotspots: VSAM↔DB2 dual persistence with mismatched keys, COMP-3 money precision, the checkpoint/restart spine, and the empty POSUPDT.cbl stub.

Contents: component inventory + dependency map, 6-R weighted scoring matrix, three-way sequencing decision matrix, 6-phase roadmap with entry/exit gates, data + service migration strategies, banking risk/governance controls, and ASCII diagrams (dependency map, disposition map, roadmap, target architecture) per the "visualize concepts" preference. Framework is written to be reusable for other mainframe-exit workloads.

A rendered PDF copy is attached to the requester in the session.

Link to Devin session: https://app.devin.ai/sessions/d4ca52d75f354735b6fef097a5856644
Requested by: @rayfan-cog

---

Devin Review

Status	Commit
⚪ Not started	—

Run Devin Review

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment, CI, and merge conflict monitoring

[devin-ai-integration]

Devin Review found 3 potential issues.

[devin-ai-integration]

🔍 DB2 schema row: count of 5 does not match the 6 objects listed

The inventory row claims 5 DB2 schema artifacts but enumerates six names (PORTFOLIO_MASTER, INVESTMENT_POSITIONS, TRANSACTION_HISTORY, POSHIST, RTNCODES, ERRLOG). The repo has exactly 5 SQL files (src/database/db2/RTNCODES.sql, PORTPLAN.sql, POSHIST.sql, db2-definitions.sql, ERRLOG.sql), so the count is per-file while the list is per-table — the two columns are measuring different things and read as an inconsistency. Also PORTPLAN.sql is not represented in the list at all.

---

Was this helpful? React with 👍 or 👎 to provide feedback.

[devin-ai-integration]

🔍 Grep-verified counts hold only when template files are excluded

I re-ran the stated greps: across src/programs the numbers match (EXEC SQL in 14, EXEC CICS in 8, VSAM ASSIGN in 25). Including src/templates/ they become 15 and 26 respectively. The COMP-3 in 7 files claim is the least reproducible: 8 files repo-wide contain COMP-3, and only 3 under programs+copybook. Worth stating the scope of the grep in the document so a reader can reproduce the inventory.

---

Was this helpful? React with 👍 or 👎 to provide feedback.

[devin-ai-integration]

📝 Info: POSUPDT is simultaneously slated for Refactor and Retire

The disposition map places POSUPDT under REFACTOR (money movement/cost basis) and POSUPDT.cbl (empty stub) under RETIRE, and §6.2 item 3 again lists POSUPDT as Refactor while item 4 calls for reverse-engineering the stub. src/programs/batch/POSUPDT.cbl is indeed 0 lines. The intent is presumably 'retire the empty artifact, refactor the behavior once reverse-engineered', but as written the two boxes contradict each other for a steering-committee audience.

---

Was this helpful? React with 👍 or 👎 to provide feedback.