Efficient Reliable Database Migration: A Legacy Innovation Story
summary
Claims a working method to lift terabytes (even hundreds of TB) from a legacy network DB into modern RDBMS by read raw disk pages and tracking old-to-new block mappings, not by slow SQL dumps. Implementation uses a quick J functional prototype, generated C and PLS mover code, binary batch serialization and sort-merge patching of pointer references. Verification uses decision-table IV&V and independent validators running invariant checks that catch real bugs. Teams run migrations on new hardware with health checks; customers accept results and only a few complex readers require manual rewrite.
AI-written summary · about
chapters (19)
- 0:00Intro
- 1:53Reality of data migration
- 2:21Case study: Critical legacy problem
- 5:38Reporting problem
- 6:42Reporting solution
- 7:45Near-term solution
- 8:22Long-term migration challenge
- 9:37Conventional migration approach
- 10:52Innovative solution
- 12:15Logical & physical model
- 13:34Essence of approach
- 15:37Binary is better
- 16:09Batch is better
- 16:43Efficient serialization
- 17:49Conservative implementation strategy to reduce risk
- 20:28Verification & validation approach
- 29:53Readers & writers?
- 32:53Recap
- 37:39Outro
related talks
39mRethinking RDBMS Data Migration
36m5 Hard-Learned Lessons from Successful Database Migration and Modernization
54mLessons learned from refactoring multi terabyte databases for the last 10 years
35mDatabase migrations without downtime
31mModernizing a megascale database
42mMigrating a database over 10TB from DB2 to PostgreSQL with zero downtime
30mAn Assessment-to-Adoption Path from Oracle, SQL Server, Sybase, DB2 onto
37mDatabase Modernization at Cloud Scale: Minimizing Risk and Disruption