Design/explore RDBMS (Oracle) to Cosmos DB NoSQL modernization assessment process/tool using GHC/LLM #9
Description
Design/pilot RDBMS (Oracle) to Cosmos DB NoSQL modernization assessment process using GHC/LLM .
(inspired by current Mongo Relational Migrator assessment tool and our own .NET App modernization assessment tool in V Studio)
Need to be able to
-
analyze RDBMS schema for Entities/relationships conversion to NoSQL based on ERD or DDL. (in case relationships missing in DDL and handled in App - ability to extract relationships from App/ORM code).
-
Identify and separate different Data Domains/groups of Entities to group together for partial migrations targeting.
-
analyze/suggest solutions for Datatypes mismatches and conversions (CLOB/BLOB to ADLS, type precisions, dates/timestamps conversion)
-
analyze/ account data volumes based on table/entity mappings (database table stats)
-
analyze/account R/W access patterns and concurrency based on top Queries, top executions ( AWR, statspack stats)
-
analyze/provide NoSQL design and sample implementations for any DB server-side Code and RDBMS objects:
- sequences
- views
- mviews
- partitioned tables (when using list/range or multi-level )
- triggers
- functions
- procedures
- packages
-analyze RDBMS/Oracle RBAC, audit and map to Cosmos DB NoSQL RBAC as part of datamodel analysis -
analyze RDBMS/Oracle external DB connections and replication options:
- dblinks
- Replication to/from
- 3rd party ETL integrations -
analyze Multi record/Multi-table transactions usage between PL/SQL DB code and Application and summarize Cosmos NoSQL DB/App changes and solution patterns to convert
-
analyze app code and summarize APP code changes based on Database conversion analysis .
OUTPUTS:
- classification of complexity of RDBMS to NoSQL (based on amount of incompatible features and their usage )
- proposed RDBMS to NoSQL Entity/Relationship mappings (NoSQL Containers, indexes and sample JSON Doc schema for each mapped to RDBMS tables/entities)
- itemize any incompatible datatypes and proposed solutions
- itemize each RDBMS specific DB structure or ServerSide PL/SQL code pattern used for solution recommendations for NoSQL conversion in both Cosmos DB as well as App Code.
- propose Alternative Cosmos DB (Mongo vCore), or Other Azure Native (AI Search, ADLS/Fabric) solutions and sync patterns for components which are not a good Cosmos NoSQL fit.
DataMigration and App conversion - PATH to PROD:
- isolated DataDomains, Stangler pattern Data and APP implementation technical example for Data+App migration.
- data pipeline for merge and sync of complex relationships in NoSQL (OFFLINE and ONLINE)