AWS RDS
- Relational Database Service is a offering for the following databases
- SQL Server
- Oracle
- mysql
- Postgres
- Aurora
- RDS databases are part of your vpc and it requires atleast two zones to be selected via DB Subnet Group.
- RDS will have a security group
- RDS supports
- single database
- Multi AZ Database
- Cluster
- RDS Support Read Replication which can be use for querying. But here the data inserted in primary will be asynchronously updated.
- RDS supports storage autoscaling minsize is 20GB
AWS RDS
-
Relational Database Service supports:
- SQL Server
- Oracle
- MySQL
- PostgreSQL
- Aurora
-
Runs inside VPC and requires at least two AZs via a DB Subnet Group.
-
Secured with Security Groups.
-
Supports deployment types:
- Single DB
- Multi-AZ Database
- Cluster
-
Supports Read Replicas (async replication, good for queries).
-
Storage Autoscaling: min size 20 GB.
-
Instance Types:
| Family / Type | Description / Use-Case |
|---|---|
| Burstable / General-Purpose (T class) | Low baseline CPU with burst credits; dev/test workloads. (db.t3, db.t4g) |
| General Purpose (M class) | Balanced CPU & memory; versatile for production. (db.m5, db.m6g) |
| Memory-Optimized (R, X classes) | High RAM per CPU; large in-memory, caching, analytics. (db.r5, db.r6g, db.x1e) |
| Compute-Optimized (C class) | High CPU per RAM; compute-heavy workloads. (db.c5) |
| Storage / I/O Optimized | Optimized for high throughput I/O workloads. |
| Legacy / Previous Gen | Older types (db.t2, db.m4, db.r4); less efficient, being phased out. |
Aurora
| Feature / Guarantee | Aurora (MySQL-compatible) | Aurora (PostgreSQL-compatible) | General Aurora Guarantees |
|---|---|---|---|
| Throughput | Up to 5× faster than MySQL. | Up to 3× faster than PostgreSQL. | Benchmarks on same hardware. |
| Read Replicas | ≤ 15 replicas; very low lag. | ≤ 15 replicas; low-latency reads. | Replica lag usually in single-digit ms. |
| Query Acceleration | Parallel Query for faster analytics. | Optimized Reads (NVMe, tiered cache) → up to 8× faster for I/O-heavy queries. | Features for large dataset performance. |
| Storage Auto-scaling | Auto-expands up to 128 TiB. | Same. | Decoupled, fault-tolerant, auto-healing, multi-AZ. |
| Availability / Durability | 6 copies across 3 AZs, auto failover. | Same. | SLA ~99.99%, Global DB for DR. |
| Compatibility | MySQL 5.6, 5.7, 8.0 (region-dependent). Some features unique to MySQL edition. | PostgreSQL-compatible with varying extensions. | Minimal migration effort. |
| Read/Write Scaling | Many read replicas; single writer; strong read scaling. Writes boosted by Aurora’s log-structured storage. | Similar, with concurrency improvements; write scaling gains less than MySQL’s. | Serverless & compute autoscaling options. |
| Latency | “Single-digit ms” replica lag; write/read speed-ups over stock MySQL. | Same, plus up to 8× read latency improvement in optimized cases. | Performance depends on workload & instance type. |
| SLA / Uptime | 99.99% uptime SLA for clusters. | Same SLA; higher availability possible with multi-region setups. | Requires Multi-AZ + replicas for SLA compliance. |
| Fault Tolerance / Replication | 6-way replication across 3 AZs; auto repair; fast failover. | Same. | Cross-region replication with Global Database for DR and global low-latency reads. |
