A. SCT

• schema conversion tool, can install on op-prem devices.
• for heterogeneous migration, only
• eg: fssrr : oracle --> SCT --> Postgres/aurora
• 

---

B. DMS

• supports : homo + hetero(with SCT)
• concept
• on prem db (large) --> offline : snowball --> aws cloud
• on-prem-db --> online : run DMS on EC2 --> aws cloud
  • Serverless option also there.

---

DMS : source

• on-prem : db ( Oracle, MS SQL Server, MySQL, MariaDB, PostgreSQL, MongoDB, SAP, DB2)
• azure : sql-db
• aws:
  • ec2 : db ( Oracle, MS SQL Server, MySQL, MariaDB, PostgreSQL, MongoDB, SAP, DB2)
  • rds
  • aurora
  • s3
  • DocumentDB

---

DMS : targets

• on-prem : db ( Oracle, MS SQL Server, MySQL, MariaDB, PostgreSQL, SAP)
• aws:
  • ec2 : db ( Oracle, MS SQL Server, MySQL, MariaDB, PostgreSQL, SAP)
  • rds
  • aurora
  • s3
  • DocumentDB
  • neptune
  • redis
  • DynamoDB
  • analytic and stream kind:
    • Kineses Data stream
    • Apache kafka
    • OpenSearch Service ?
    • Redshift ?

---

network perspective

• Transferring large amount of data into AWS, from network perspective
• Over the internet / Site-to-Site VPN:
• Over direct connect1Gbps
• Over Snowball offline

---

DMS : Continuous replication

• use SCT and DMS
• also enable multi-zone on target db

---

demo

// 1. DMS - `enpoints`
- endpoint-1 to connecto source DB
- endpoint-2 to connecto target DB

// 2. DMS - `replication-instance`
  - choose starting point:
    - discover and access
        - give roadmap, 
        - generate path in few hrs
    - convert
        - for hetero, use sct
        - takes weeks
    - migrate
        - choose:
            - instance-based (mamge/admins ec2) **
                - provision ec2 intance
                - DMS version
                - single-az/multi-az : az replication
                - storage for ec2
                - network
                - maintence window
            - serverless (no admin, migrated + replicates)
                - EASY

// 3. DMS `replication-task`
- create task-1
    - choose source endpoint
    - choose target endpoin
    - choose replication instance 
        - task setting : { json } - update it

---

scenario

• 1 RDS(mySQl) --> migrate(no DMS) --> Aurora(engine:mySQL)

• option-1: RDS > rds::snapshot > s3 > restore > Aurora

• option-2: aurora > create read-replica from RDS > promote it.

• 2 ext-mysql --> migrate(no DMS) --> Aurora(engine:mySQL)

• option-1 : ext-mysql > percona-mysql-tool:backup > s3 > restore > Aurora
• option-2 : ext-mysql > dump-util > restore dumps > aurora : SLOW (no s3)

• 3 RDS(postgres) --> migrate(no DMS) --> Aurora(engine:postgres)

• option-1: RDS > rds :snapshot > s3 > restore > Aurora

• option-2: aurora > create read-replica from RDS > promote it

• 4 ext-postgres --> migrate(no DMS) --> Aurora(engine:postgres)

• option-1 : RDS > backup-tool:backup > s3 > import(aws_se_aurora extension) > Aurora