Multi-AZ vs Multi-Region vs Read replicas - Amazon RDS - AWS Certification Cheat Sheet

Let’s get a quick overview of Multi-AZ vs Multi-Region vs Read replicas for Amazon RDS from an AWS certification perspective.

You will learn

  • What are Multi-AZ Deployments?
  • What are Read replicas?
  • How do you choose between Multi-AZ deployments vs Multi-Region vs Same-Region Read replicas?

Table of Contents

AWS Certification Study Material and Notes - 25 PDF Cheat Sheets

Each cheat sheet contains:

  • FAQs and Tutorials with 5-20 slides
  • Study Material and Notes to quickly review and prepare for certification exam
  • Certification Exam Tips
  • Certification and Interview Scenario Questions

Multi-AZ Deployments

  • Standby created in a different AZ
  • Synchronous replication (strong consistency)
  • Enhances durability, availability and fault tolerance of your database
  • Multi-AZ makes maintenance easy
    • Perform maintenance (patches) on standby
    • Promote standby to primary
    • Perform maintenance on (old) primary
  • Avoid I/O suspension when data is backed up (snapshots are taken from standby)
  • No downtime when database is converted to Multi AZ
    • Increased latency until standby is ready
  • Not allowed to connect to standby database directly
    • For example: Standby CANNOT be used to serve read traffic
    • Standby increases availability but does not improve scalability
  • Automatic failover to standby if master has problems (compute, storage or networking)
    • CNAME record flipped to standby
    • Database performance issues (long running queries or deadlocks) will NOT cause a failover
  • (Good Practice) Use DNS name of database in applications configuration

Read Replicas

  • Support read-heavy database workloads - reporting and data warehousing
  • Can be in same or different AZ or different Region
  • Your apps can connect to them
  • Create read replica(s) of a read replica
  • Uses asynchronous replication
    • Provides eventual consistency (from replica)
    • For higher consistency, read from master
  • Need to be explicitly deleted (Not deleted when database is deleted)
  • (Mandatory) Enable automatic backups before you can create read replicas
    • Set Backup Retention period to a value other than 0
  • Reduce replication lag by using better compute and storage resources
  • Maximum no of read replicas:
    • MySQL, MariaDB, PostgreSQL, and Oracle - 5
    • Aurora - 15
    • SQL Server does not support read replicas

RDS Multi-AZ vs Multi-Region vs Read replicas - Certification Questions and Scenarios

Feature Multi-AZ deployments Multi-Region Read Replicas Multi-AZ Read replicas
Main purpose High availability Disaster recovery and local performance Scalability
Replication Synchronous (except for Aurora - Asynchronous) Asynchronous Asynchronous
Active Only master (For Aurora - all) All read replicas All read replicas


Certification - Recommended Reading

Elastic Load Balancer (ELB) vs Network Load Balancer (NLB) - A Difference

Elastic Beanstalk vs Cloudformation vs Opswork vs Codedeploy - A Difference

EBS HDD Storages - A Difference

AWS Shield vs AWS WAF vs AWS Macie - Protect Resources and Data Cheat Sheet

Managing Multiple AWS Accounts - Organizations, Trusted Advisor and more Cheat Sheet

Amazon CloudWatch - Logs, Events, Alarms and Dashboards Cheat Sheet

AWS CloudTrail vs Config vs CloudWatch Cheat Sheet

EBS SSD Storages - A Difference

NAT Gateway vs Instance - A Comparison

AWS Security Groups and Network Access Control List - A Comparison Cheat Sheet

Free Videos - Get Started with AWS Certification

Amazing AWS Certification Courses

Get our amazing courses pursued by thousands of learners