AWS Classroom notes 24/Jan/2024

AWS RDS Contd

• Instance types: this is vm size on which database will be running
  • Burstable: this is equivalent to t-series in ec2 instance
  • Memory optimized: these machines have more ram to cpu ration and it is r-series
  • Serverless: Here we will be charged per millisecond based on cpu and ram consumption (This option is enabled for aurora)
  • Standard classes
• Cost Estimate
  • mysql
  • postgres
  • oracle
  • microsoft sql server
  • aurora mysql
  • aurora postgres
  • ibm db2

mysql db.m1.xlarge 4 cpu 15 GB RAM 100 GB mutli-az = 701 $ 
postgres db.m1.xlarge 4 cpu 15 GB RAM 100 GB multi-az = 782 $
oracle enterprise db.m5.xlarge 4 cpu 16 GB RAM 100 GB single-az = 310 $ + license costs 
Sqlserver  db.m5.xlarge 4 cpu 16 GB RAM 100 GB single-az = 449 $ + license costs 
aurora mysql db.r3.xlarge 4 cpu 30.5 GB RAM 100 GB multi-az = 477.73 * number of nodes
aurora postgres db.r5.xlarge 4 cpu 30.5 GB RAM 100 GB multi-az = 477.73 * number of nodes

• Storage: This uses disks i.e. ebs volumes, generally we have a general purpose volume and provisioned iops. Here we have an option for storage autoscaling which will automatically increase disk capacity till the maximum extent provided
  

Single-AZ Deployments

• In this case your database will be created in one AZ from dbsubnet group.
• The database created needs to handle both read and writes.
• If this database fails, we need to recreate or restore

Muti-AZ Deployments

• Multi AZ creates multiple db servers and supports automatic failover. the endpoint will not change post failover

Read Replica

• AWS create a replication of database where only read is allowed.
• This can be implemented with Single AZ as well as multi-az
• Read replica can be created in other region as well.