Section 5 - RDS, Aurora & ElastiCache

Relational Database Service is a managed service for databases using SQL as a query language
Create databases such as Postgres, MySQL, MariaDB, Oracle, Microsoft SQL Server, IBM DB2 and Aurora (AWS proprietary database)
Better to use RDS over deploying a DB on EC2
- Automated provisioning, OS patching
- Continuous backups
- Dashboards
- Read replicas for improved performance
- Multi AZ setup for disaster recovery
- Scaling
- EBS storage

Can have up to 15 within AZ, Cross AZ or Cross Region
Replication is asynchronous so reads are eventually consistent
Replicas can be promoted to their own DB
Applications must update the connection string to leverage read replicas
There's a network cost when data goes from one AZ to another, for RDS Read Replicas within the same region, there is no fee
Use cases:
- Production database taking on a normal load
- Run a reporting app to run analytics
- Want a replica to run new workload
- Want the production app to be unaffected
- Read replicas are used for SELECT kind of statements, not INSERT, UPDATE and DELETE

Synchronous replication, increased availability
Failover incase of loss of AZ, network, instance or storage failure
No manual intervention in apps, not used for scaling
Moving from a Single-AZ to Multi-AZ setup has zero downtime
- Takes a snapshot and restores the new DB in a new AZ
- Synchronization is established between the two databases

Access to underlying databases and OS
- Configure settings
- Install patches
- Enable native features
- Access underlying instance using SSH or SSM Session Manager
Better to de-active automation mode to perform customization, also take a DB snapshot prior

Supports Postgres and MySQL
5 and 3 times more performance over MySQL on RDS and Postgres on RDS respectively
Storage automatically grows in increments of 10GB up to 128TB
Up to 15 replicas, the replication process is faster than MySQL
Failover is instantaneous
Cost about 20% more than RDS, but more efficient

6 copies of data across 3 AZ, 4 are needed for writes and 3 are needed for reads
Master Aurora instance takes writes
Automated failover for master in less than 30 seconds
Support for Cross Region Replication

Add ML-based predictions to your apps via SQL
Simple and secure integration between Aurora and AWS ML services (SageMaker and Comprehend)
Useful for fraud detection, ads targeting, sentiment analysis, product recommendations, etc.

RDS
- Automated:
  - Daily full backup of DB
  - Transaction logs are backed up every 5 minutes
  - Restore to any point in time
  - 1 to 35 days of retention
- Manual:
  - Triggered snapshot by user
  - Retention of backup for as long as you want
Aurora
- Automated:
  - 1 to 35 days (can't be disabled)
  - Point-in-time recovery in that time-frame
- Manual:
  - Triggered snapshot by user
  - Retention of backup for as long as you want

Restoring a backup or snapshot creates a new DB
Restoring MySQL RDS DB from S3
- Create backup of on-premises DB, store it on S3 and restore the backup file onto a new RDS instance running MySQL
Restoring MySQL Aurora cluster from S3
- Create backup of on-premises DB using Percona XtraBackup, store backup file on S3 and restore it onto a new Aurora cluster running MySQL

Create new DB Cluster from an existing cluster, faster than snapshot and restore
Uses copy-on-write protocol
Fast and cost-effective
Useful for creating a staging database from a production database without impacting the production database

At-rest encryption: Use AWS KMS defines at launch time
In-flight encryption: TLS-ready by default, use TLS Root certificates client-side
IAM authentication: IAM roles to connect to DB
Security groups: Control network access to database
No SSH available except on RDS custom
Audit logs can be enabled and sent to CloudWatch Logs for longer retention

Enables apps to pool and share established DB connections
Serverless, auto scaling, high availability with Multi-AZ setup
Improves DB efficiency by reducing stress on DB resources and minimizing open connections
Reduced RDS and Aurora failover time by up to 66%
Proxy only accessible from VPC, not public

In-memory database with high performance and low latency
AWS managed service used to get managed Redis or Memcached
Reduce read intensive workloads, make apps stateless
Using ElastiCache involves heavy application code changes
- Apps query ElastiCache, if it's not available then query RDS and store in ElastiCache
- Cache must invalidate stale data

Redis
- Multi-AZ with auto-failover
- Read Replicas to scale reads and high availability
- Backup and restore features
- Support sets and sorted sets
Memcached
- Sharding: Multi-node for partitioning of data
- No replication
- Not persistent
- Serverless backup and restore
- Multi-threaded architecture

Supports IAM authentication for Redis
IAM policies only used for AWS API-level security
Redis AUTH:
- Set a password/token upon creating a cluster
- Support SSL in-flight encryption
Memcached:
- Supports SASL-based authentication

Lazy loading: Cache all read data, but data can become stale
Write through: Adds or update data in cache when written to DB (solves stale data)
Session store: Store temporary session data using TTL features