Matching the performance of a high-transaction, high-value workload to the right database is critical. Different database products use different techniques to increase performance, and different techniques can lead to extra work and extra cost for customers down the line.
Aurora and MySQL are scale-up database products. Scale-up databases are designed to run on a single instance (or node), and customers can increase performance by migrating to a larger, faster and typically more expensive instance. Once scale-up databases are running on the largest available hardware, the process of increasing performance moves to adding read slaves and/or sharding — techniques which are expensive to implement and maintain, add fragility and do little to scale the write performance of the database workload.
ClustrixDB is a true scale-out database. Scale-out databases are designed to run on a cluster of affordable instances (typically three or more) and provide performance and reliability appropriate for use with high-transaction, high-value workloads. Scale-out databases are immune to any single point of failure and can scale both reads and writes in a near linear fashion. And the ClustrixDB Rebalancer feature is constantly evaluating and balancing the distribution of data across the cluster to ensure optimal performance.
Scale-out databases are particularly well suited for mixed workloads where there is normal online transactional processing (OLTP) running concurrently with report creation and even some business analytics. The increased demands of a mixed workload have traditionally overwhelmed scale-up databases like MySQL and Aurora, but they are not a problem for ClustrixDB.
|Scales Reads||Yes, through read slaves||Yes, through replicants (read slaves)||Yes. No read slaves or replicants required.|
|Automatic rebalancing of data across all instances||No||No||Yes|
|Appropriate for mixed workloads||No||No||Yes|
ClustrixDB is designed to scale and automate common tasks that require human workarounds for MySQL and Aurora.
Scale-out databases like ClustrixDB are transparent to enterprise applications. Regardless of the number of nodes in the cluster, applications see only a single, large, fast database.
ClustrixDB is the only drop-in replacement for MySQL that can scale-out both reads and writes. With ClustrixDB you can get 10x the maximum performance of Aurora (or MySQL) without the complexity and fragility of read slaves, sharding or modifying your enterprise applications.
Read this white paper that provides a distributed relational database architecture overview and explains how ClustrixDB scales reads and writes.