Very interesting presentation in Amazon Event “From Mainframe to Microservices – Vanguard’s Move to the Cloud”. Following parts can be used as Migration Patterns: Different ways of how to move from Mainframe to Microservices:
- Re-host
- Reengineer
- Refactor
- Rewrite
- Batch job migration
- Data analytics enablement
- Data-driven augmentation
- Re-platform with Linux and Java
- Repurchase
- Retire
- Combination of all
This presentation also shows the Cloud Data Architecture at Vanguard. Generally this is how we can build our own database for our micrsoservices at the cloud and at the same time still use the data from DB2 mainframe on-premise as well.
Advantages using this data architecture:
- Each microservice reads data from its own local database.
- DB2 mainframe database can still be used from other Host / Cobol batches.
- DB scalability
Disadvantages:
- You cannot write to the database in case you have a network problem. There is an extension to this Cloud Data Architecture by using buffered write or asynchronous write.
- The disadvantage using buffered write is the eventual consistency of the local database.
You shoud take a look at the presentation if you need to move to cloud and you want to do it step by step with Strangler Pattern.
| Published on Java Code Geeks with permission by Lofi Dewanto, partner at our JCG program. See the original article here: Moving Host Cobol Batches and Monolith Webapps to Cloud and Microservices Opinions expressed by Java Code Geeks contributors are their own. |
