Differences between RDS Multi-AZ and Read Replica, explained for beginners
Organizes the differences between Amazon RDS Multi-AZ deployments and Read Replicas — easily confused — across availability, read performance, failover, and exam-relevant points.
Services compared
Click a service name to review it on the term detail page.
Bottom line
RDS Multi-AZ and Read Replica both involve copying a database.
However, the purpose is different.
- Use Multi-AZ when the goal is fault tolerance and higher availability
- Use Read Replica when the goal is to scale read performance
This distinction matters a lot on AWS exams.
Comparison table
| Item | Multi-AZ | Read Replica |
|---|---|---|
| Main purpose | High availability and fault tolerance | Better read performance |
| Replication | Synchronous replication by default | Asynchronous replication |
| Read distribution | The standby is not readable in the standard setup | The replica is readable |
| Failover | Automatic failover | Not intended for automatic failover |
| Placement | Standby in a different AZ | Same Region, or a different Region |
| Writes | Primary only | Primary only |
| Reads | Read from the primary | Read from the replica |
| Exam keywords | high availability, failover, standby | read scaling, read-heavy workload |
| Cost | Adds the cost of the standby | Adds the cost of each replica |
Beginner-friendly explanation
Multi-AZ is the mechanism that prepares the database for a failure on the main DB.
Primary DB
↓ synchronous
Standby DBIn normal operation, the application talks to the Primary DB.
When a failure happens, traffic switches over to the Standby DB.
Read Replica is a copy of the DB used to absorb read traffic.
Primary DB
↓ asynchronous
Read ReplicaThe application can route writes to the Primary DB and reads to the Read Replica.
Exam points
Cases where Multi-AZ is the likely answer
- Improve database availability
- Prepare for an AZ failure
- Automatic failover is required
- Add fault tolerance to a production DB
- Shorten the RTO
Cases where Read Replica is the likely answer
- A lot of read access
- Move reporting or analytics queries away from the production DB
- Increase read throughput
- Place a read-only DB in a different Region
How to choose in practice
Fault tolerance for a production DB
Use Multi-AZ.
Application
↓
RDS Primary
↓ synchronous
RDS StandbyIn production, Multi-AZ is commonly used to avoid a single point of failure on the DB.
A service with heavy read load
Use Read Replica.
Write request → Primary DB
Read request → Read ReplicaIt fits read-heavy workloads such as product browsing on an e-commerce site, article reading on a media site, or dashboard aggregations in an admin console.
When both high availability and read performance are needed
Combine Multi-AZ and Read Replica.
Primary DB
├─ Standby DB
└─ Read ReplicaThat said, the cost also goes up, so this combination is not used in an indie MVP by default.
Common misunderstandings
Misunderstanding 1: treating Read Replica as the first answer for fault tolerance
The main purpose of Read Replica is to improve read performance.
The first answer for fault tolerance is Multi-AZ.
Misunderstanding 2: expecting Multi-AZ to improve read performance
In the standard RDS Multi-AZ DB instance setup, the standby is not used to serve reads.
To scale read performance, consider Read Replica.
Misunderstanding 3: using RDS in an indie MVP
This project does not use RDS in the MVP.
The reason is that the fixed cost ramps up quickly, and if the only need is to store inquiry submissions, Amazon DynamoDB is enough.
Related terms
Related questions
Summary
The clearest way to remember RDS Multi-AZ and Read Replica is by purpose.
- Multi-AZ: fault tolerance and high availability
- Read Replica: better read performance
On the exam, when "high availability" appears in the question, suspect Multi-AZ; when "read-heavy workload" appears, suspect Read Replica.
What to learn next
After understanding the differences in this comparison, review related terms and practice questions. AWS exams often test how to choose between similar services.