RDS MySQL provides multiple editions to suit different business scenarios. This topic helps you understand the differences between these editions and compare their parameters, to help you quickly select the most suitable configuration.
Quick purchase
Priority | Recommended combination | Rationale |
Cost-sensitive | Low cost with a single-node architecture. Supports automatic failure recovery and automatic start and stop. Dynamically scales resources based on instance load, with an SLA of 99.50%. | |
Cost and availability | Cost-effective: Secondary nodes are accessible and scalable. The high-performance cloud disk supports I/O burst (with a free quota), Buffer Pool Extension (BPE) (free), and data archiving to reduce cold data storage costs. High availability: ensures lossless failover for availability zone failures and fast recovery from regional failures. It also supports a cross-cloud disaster recovery architecture, with an SLA of up to 99.99%. | |
Performance and availability | High performance: Delivers very low I/O latency (10–30 microseconds). You can add read-only instances to handle high read traffic. High availability: ensures lossless failover for availability zone failures and fast recovery from regional failures. It also supports a cross-cloud disaster recovery architecture, with an SLA of up to 99.99%. |
Key configuration options
Billing method (affects instance cost and elasticity)
Option | Key features |
Pay in advance to reserve resources and receive significant discounts. | |
Pay for what you use on an hourly basis. For some instance types, you can stop the instance to suspend billing. | |
Pay for what you use on an hourly basis. Resources are dynamically scaled based on workload, and it supports automatic start and stop of instances. | |
Changing billing methods: You can switch between subscription, pay-as-you-go, and serverless billing methods. | |
Edition (affects instance availability, performance, elasticity, and cost)
Option | Architecture | Horizontal scaling | Disaster recovery | Failover | Read/write splitting | Data consistency |
Single-node | Not supported | Weak | Not supported | Not supported | Weak | |
One primary, one secondary (Secondary node is inaccessible) | Secondary node is not scalable Can add read-only instances | multi-AZ deployment Cross-region disaster recovery | Automatic | Supported | Strong | |
One primary and one or more secondaries (Secondary nodes are accessible) | Scalable secondary nodes (Up to 8) | multi-AZ deployment Cross-region disaster recovery | Automatic | Supported | Stronger (Group Replication ensures RPO=0) | |
Changing editions: You can upgrade the Basic Edition to the High-availability or Cluster Edition. You can also upgrade the High-availability Edition to the Cluster Edition. | ||||||
Storage type (affects instance performance, cost, and scaling speed)
If you require low I/O latency, choose a high-performance local disk. It provides low latency (10 to 30 microseconds), and its storage cost is lower than that of ESSD PL1.
If low I/O latency is not a key requirement, choose a high-performance cloud disk. Its baseline performance is between that of PL1 and PL2, and its storage cost is comparable to that of PL1. You can enable Buffer Pool Extension (BPE) and I/O burst for significant improvements in IOPS and throughput, and use data archiving to reduce storage costs.
Option | I/O latency | Backup speed | Scaling speed | Max IOPS | Max throughput | Storage cost | |
Low | Medium | Medium, causes a brief connection interruption | Determined by the instance type | Lower than PL1 | |||
cloud disk | ESSD PL1 | Medium | Fast | Fast, with no connection interruption | Baseline | ||
(Baseline performance) | Higher than PL1, lower than PL2 I/O burst can be enabled for sustained increases in performance. | ~PL1 | |||||
ESSD PL2 | ~2x | ~2x | ~2x | ||||
ESSD PL3 | ~20x | ~11x | ~4x | ||||
ESSD PL0 | ~0.2x | ~0.5x | ~0.5x | ||||
Being phased out | |||||||
| |||||||
The actual maximum IOPS and throughput of an instance are affected by multiple factors, including the storage type, instance type, and storage capacity. For more information, see Primary RDS MySQL instance types and Read-only instance types.
Product type (affects instance cost, performance, and supported MySQL versions)
Option | Storage type | MySQL version | Supported regions and types | Max instance type | |
Standard Edition (X86) | cloud disk, high-performance local disk | 8.4/8.0/5.7/5.6/5.5 | All | 104-core 768 GB | Same |
Yitian Edition (ARM) | cloud disk | 8.4/8.0/5.7 | Some regions and instance types | 64-core 512 GB | |
You can evaluate the performance and pricing of the Yitian Edition online for free. | |||||
Changing product types: You can switch between the Standard and Yitian editions. | |||||
Feature comparison
In the table, ✔️ indicates that the feature is supported, and ❌ indicates that it is not.
PL0, PL1, PL2, and PL3 indicate different performance levels of ESSD cloud disks.
Criteria | Basic Edition | Cluster Edition | High-availability Edition | |||
Cloud disk | Cloud disk | Cloud disk | High-performance local disk | |||
Supported MySQL versions | 8.4/8.0/5.7 | 8.4/8.0/5.7/5.6/5.5 (New purchases of v5.5 are discontinued) | ||||
Availability | multi-AZ deployment | ❌ | ✔️ | ✔️ | ✔️ | |
❌ | ✔️ | ✔️ | ✔️ | |||
❌ | ✔️ | ✔️ | ✔️ | |||
Elasticity | Scaling specifications: Takes minutes and causes a brief connection interruption. Scaling storage up: Takes seconds and does not interrupt service. Scaling storage down: The duration depends on the data volume and causes a brief connection interruption. | Time varies based on data volume. If no host migration is required, scaling takes minutes. Causes a brief connection interruption. | ||||
❌ | ✔️ | ✔️ | ✔️ | |||
✔️ | ❌ | ✔️ | ❌ | |||
Performance | QPS / TPS | Refer to the RDS MySQL Performance White Paper to select an instance type, and then adjust the instance configuration based on your actual workload after migration. | ||||
high-performance local disk: 10–30 microseconds; ESSD PL1/PL2/PL3/high-performance cloud disk (baseline): 200 microseconds ESSD PL0: 300–500 microseconds; SSD cloud disk: 500–2,000 microseconds | ||||||
PL0: 10,000; PL1: 50,000; PL2: 100,000; PL3: 1,000,000 high-performance cloud disk: Baseline of 72,400, bursts up to 1,000,000 SSD cloud disk: 25,000; high-performance local disk: Depends on the instance type. | ||||||
PL0: 180 MB/s; PL1: 350 MB/s; PL2: 750 MB/s; PL3: 4,000 MB/s high-performance cloud disk: Baseline of 700 MB/s, bursts up to 4,096 MB/s SSD cloud disk: 300 MB/s. High-performance local disk: Varies with the high-performance local disk instance types. Note The actual maximum IOPS and throughput of an instance are affected by multiple factors, including the storage type, instance type, and storage capacity. For more information, see Primary RDS MySQL instance types and Read-only instance types. | ||||||
PL0: 32,000 GB; PL1/2/3: 64,000 GB; SSD cloud disk: 6,000 GB high-performance local disk: 3,000–16,000 GB, depends on the instance type. | ||||||
Standard Edition | 8-core 32 GB | 104-core 768 GB | 104-core 768 GB | 90-core 720 GB | ||
Yitian Edition | 2-core 4 GB | 64-core 512 GB | 32-core 128 GB | ❌ | ||
Cost | Example: A 1-month subscription for a 2-core 8 GB general-purpose instance with 50 GB of storage in the China (Hangzhou) region. The Basic Edition is single-node. The High-availability and Cluster editions are two-node (one primary, one secondary). The cloud disk type is high-performance cloud disk. (The following prices are list prices for reference only. The actual price on the product page prevails.) | |||||
Standard Edition price | CNY 240/month | CNY 686/month | CNY 685/month | CNY 940/month | ||
Yitian Edition price | ❌ | CNY 470/month | CNY 469/month | ❌ | ||
Security | storage encryption | ❌ | ✔️ | ✔️ | Supported only on v5.6 and later | |
✔️ | ✔️ | ✔️ | ❌ | |||
in-transit encryption (SSL) | ✔️ | ✔️ | ✔️ | ✔️ | ||
✔️ | ✔️ | ✔️ | Supported only on v5.7, v8.0, and v8.4 | |||
✔️ | ✔️ | ✔️ | ✔️ | |||
❌ | ✔️ | ✔️ | ✔️ | |||