TXlTUUwgb24gQXp1cmUgUGVyZm9ybWFuY2UgQmVuY2htYXJrIOKAkyBTY2FsZUdyaWQgdnMuIEF6 dXJlIERhdGFiYXNl

Microsoft Azure is one of the most popular cloud providers in the world, and a natural fit for database hosting on applications leveraging Microsoft across their infrastructure. MySQL is the number one open source database that’s commonly hosted through Azure instances. While Microsoft offers their own Azure Database product, there are other alternatives available that may be able to help you improve your MySQL performance. In this blog post, we compare Azure Database for MySQL vs. ScaleGrid MySQL on Azure so you can see which provider offers the best throughput and latency performance. We measure latency in ms 95th percentile latency.

At a glance - TLDR

Read-Intensive Workloads During Read-Intensive Workloads, ScaleGrid manages to achieve up to 3 times higher throughput and averages 66% better latency compared to Azure Database. Read now Balanced Workloads With an average of 150% better throughput, and less than one third of the latency, ScaleGrid outperforms Azure Database for Balanced Workloads. Read now Write-Intensive Workloads Scalegrid achieves up to 4 times higher throughput and averages 64% lower latency across all number of threads compared to Azure Database for MySQL. Read now

Just getting started? Check out The Best Way to Host MySQL on Azure Cloud post to learn more about optimizing your cloud database deployment.

MySQL Azure Performance Benchmark

In this benchmark report, we compare MySQL hosting on Azure at ScaleGrid vs. Azure Database for MySQL across these three workload scenarios:

• Read-Intensive Workload: 80% reads and 20% writes
• Balanced Workload: 50% reads and 50% writes
• Write-Intensive Workload: 20% reads and 80% writes

We measure the MySQL throughput and latency performance, and measure throughput in terms of queries per second (QPS) and latency in terms of 95th percentile (ms). Check out our Benchmark Configuration section below the performance report to see how these tests were configured.

MySQL Read-Intensive Performance

MySQL read-intensive workloads are workloads that are typically dominated by read operations, such as SELECT. So, a read-intensive workload would be one that searches the database more often vs. writing to it. Percona has a great post on read-intensive vs. write-intensive workloads where you can learn more.

Let’s take a look at the throughput and latency performance of ScaleGrid MySQL vs. Azure Database for MySQL:

Throughput

Latency

MySQL Balanced Workload Performance

Balanced workloads leverage roughly equal amounts of read operations and write operations.

Throughput

Latency

MySQL Write-Intensive Performance

While read operations are searching from the database, write operations are those that save to or modify the database, such as INSERT, UPDATE or DELETE queries. Write-intensive workloads are typically more expensive than read-intensive workloads as the operations consume more resources. Let’s compare ScaleGrid vs. Azure Database throughput and latency performance across MySQL write-intensive workloads.

Throughput

Latency

As we can see from the above reports, ScaleGrid helps you significantly increase your throughput and decrease your latency against Azure Database deployments for MySQL across read-intensive, write-intensive and balanced workloads scenarios. To learn more about how these two providers compare across features, check out the ScaleGrid vs. Azure Database MySQL page.

Benchmark configurations

Let’s take a look at the configurations we used in the performance benchmark:

Configuration Comparison

We designed the configuration using the most comparable plans offered between ScaleGrid and Azure Database. There are only minor differences and the costs are the same for the two providers:

Sysbench Configuration

MySQL Server Configuration

Note that at ScaleGrid, you have advanced options to customize your MySQL configurations. This allows you to customize your deployment to your application needs and optimize for performance. Unfortunately, Azure Database does not allow for customization of some of the parameters, so you are stuck with the default configurations available through their platform.

For example, we were able to customize the InnoDB log file size to 1GB whereas for Azure Database, this cannot be modified and has a default value of 268M. Learn more about the limitations in Azure Database for MySQL.

There are many other advantages of using Scalegrid MySQL for Azure, check out the full comparison on our ScaleGrid vs. Azure Database - MySQL Hosting page.