Database Benchmark on AWS: ClustrixDB Blows Aurora Away

Database Benchmark: 10X Higher Throughput, Faster Response for High-Transaction, High-Value Workloads

San Francisco, Calif. – Feb 18, 2016 – Clustrix, the provider of the first scale-out database designed to meet the elastic scaling requirements of high-transaction, high-value Web applications, announced today the results of its database benchmark test, which compared the performance of ClustrixDB to Amazon Aurora running in AWS. The results demonstrate that ClustrixDB is the top choice for companies that require or anticipate high growth, and will be placing tremendous demands on mission-critical databases, such as those used in ad tech, gaming, e-commerce, web and mobile businesses. Benchmark results will be published and discussed on February 25, 2016 at their live webinar, “Benchmark Showdown: Which Relational Database is the Fastest on AWS?”

“While AWS Aurora offers modest improvements over MySQL, companies that place high demands on their database will eventually hit a point at which they begin experiencing performance problems,” said Mike Azevedo, President and CEO, Clustrix. “ClustrixDB’s unique architecture enables it to achieve MySQL-compatibility and scalability on a much greater level.”

10x the performance of Aurora

While Aurora made headlines with 5x performance improvements over MySQL, the database benchmark shows that ClustrixDB can achieve performance beyond 10x that of Aurora. ClustrixDB consistently delivers the fastest database response times and transaction throughput over Aurora running in AWS.

“Response time can be kept at sub-20ms rates—the threshold at which busy sites begin to experience performance issues,” said Dave Anselmi, Director of Product Management, Clustrix. “Aurora’s response time gets worse as load increases—you may not notice it until you get to a certain level of performance, but for companies that anticipate rapid growth, this is going to be a serious issue.”

Scale out by merely adding nodes, flex up and down to seasonal demand

Aurora and other MySQL replacements run on a single server, and once you reach capacity you experience performance problems. These limitations can only be addressed by migrating to larger hardware—which is expensive—or techniques like read slaves, multiple masters and shards, which add complexity and fragility to applications.

By contrast, ClustrixDB users achieve increases in performance by merely adding server instances/nodes. “No matter how big you’re running, ClustrixDB is seen by your applications as one large, fast MySQL-compatible database,” continued Anselmi. “You’ll never outgrow ClustrixDB.” The ability to add or subtract nodes with just a few clicks—along with Clustrix’s flexible licensing and pricing mode—also enables companies to flex up and down to seasonal spikes, so that they avoid paying for capacity they don’t need.

Drop-in replacement for MySQL

Companies that are outgrowing MySQL can easily move to ClustrixDB without missing a beat through a well-documented, step-by-step process. As a drop-in replacement for MySQL, ClustrixDB is compatible with the existing applications, people and processes that companies already have in place, and doesn’t require replatforming or hiring new staff.

Scale reads and writes to increase and maximize performance

ClustrixDB is also unique in its ability to scale both reads and writes. For solutions that create significant ‘write’ workloads to the database (such as ad tech, gaming and e-commerce, etc.), ClustrixDB can increase the ability to process high transaction workloads without modifying applications.

Pricing comparable to the cost of Aurora

Unlike Aurora or MySQL, ClustrixDB was designed to run on mainstream nodes like the c3.2xlarge, which will cost less than 50% of the price of Aurora’s hardware with similar configurations. “We are flat out a better designed and more efficient DBMS,” said Azevedo. “We don’t need near the memory and processing that Aurora does to deliver better performance, at a lower cost.”

See a more in-depth comparison of Aurora and ClustrixDB.