Exploring AWS !!

Day 37:

Amazon RDS, RedShift and Aurora — Overview:

RDS (Relational Database Service):

• managed database service for databases using SQL as query language.
• Allows you to create database in cloud that are managed by AWS
• supports: Postgres, MySQL, MariaDB, Oracle, MS SQL server, Aurora (AWS proprietary database) Advantages over using RDS v/s deploying Database on EC2:
• RDS is a managed service
• Automated provisioning, OS patching
• Continuous backups and restore to specific timestamp (Point-In-Time-Restore)
• Monitoring dashboards
• Multi-AZ setup for Disaster recovery
• Maintenance window for upgrades
• Scaling capability (vertical and horizontal)
• Storage backed by EBS
• Can’t SSH into instance

Amazon RedShift:

• user columnar data storage
• unlike SQL, data is not stored in rows but in columns, which is ideal for data warehousing.
• requires fewer reads
• best suited for online analytical processing
• does not require any indexes or materialized views.
• cluster — combination of set of nodes, leader node and one or more compute node. Leader node received the queries from an application, parses queries and makes an executable plan.
• 2 types of nodes — Dense Storage (DS) -> Storage optimized, Dense Compute -> Compute Optimized
• Nodes are scalable
• Single node has 160 GB storage and in a cluster we can have 128 compute nodes.

Amazon Aurora:

• proprietary technology from AWS (not open sourced)
• Postgres and MySQL are both supported as AuroraDB (means your drivers work as if Aurora was a Postgres or MySQL database)
• Aurora is AWS cloud optimized and claims 5x performance improvement over MySQL and 3x improvement over PostgreSQL on RDS.
• storage automatically grows increments of 10 GB to 128 TB.
• It can have 15 replicas while MySQL has 5 & replication is faster.
• failover in Aurora is instantaneous, its High availability native.
• Aurora costs more than RDS (20% more) but is more efficient.