Decide what instance class to use - aws_mainframe_modernization_service - Latest

AWS Mainframe Modernization Data Replication for IBM z/OS

Product type
Software
Portfolio
Integrate
Product family
Connect
Product
AWS Mainframe Modernization > AWS Mainframe Modernization Service
Version
Latest
ft:locale
en-US
Product name
AWS Mainframe Modernization
ft:title
AWS Mainframe Modernization Data Replication for IBM z/OS
Copyright
2025
First publish date
2000
ft:lastEdition
2025-02-10
ft:lastPublication
2025-02-10T15:55:15.122000

To help determine which replication instance class might work best for you, let's look at the change data capture (CDC) process that AWS mainframe modernization service's data replication feature with Precisely uses.

Let us assume that you are running a CDC task (ongoing replication). In this case, the replication job has its own publisher and engine, which stores metadata and other information. The reading of transaction logs on the source is kicked off as soon as the replication job on the AWS EC2 instance is started and starts a transactional apply process on the target instance.

During the ongoing replications phase, the replication process uses two stream buffers, one each for incoming and outgoing data. The replication process also uses another memory buffer to keep track of the changes. Following are two important uses of this component (which has others):
  • It tracks all transactions and makes sure that it forwards only relevant transactions to the outgoing buffer.
  • It makes sure that transactions are forwarded in the same commit order as on the source.

As you can see, there are three important memory buffers in this architecture for CDC. If any of these buffers experience memory pressure, the overall replication can have performance issues that can potentially cause failures.

When you plug heavy workloads with a high number of transactions per second (TPS) into this architecture, you can find the extra memory provided by R5 and R6i instances useful. You can use R5 and R6i instances to hold a large number of transactions in memory and prevent memory-pressure issues during ongoing replications.