Another Scale-up Relational Database . . . Really?


Amazon today announced Aurora, a MySQL compatible RDBMS with 5x the performance of MySQL. Just what we need, another scale-up SQL database. High availability comes via fail-over of replicated instances with at least 2 copies of the original data. In this architecture they claim to easily and seamlessly scale storage:

“As your business grows, Amazon Aurora will scale with you. You won’t need to take your application off line in order to add storage. Instead, Amazon Aurora will add storage in 10 GB increments on as as-needed basis, all the way up to 64 TB. Baseline storage performance is rapid, reliable and predictable—it scales linearly as you store more data, and allows you to burst to higher rates on occasion.”

But what about overall performance? As with MySQL, any scale-up database, including Aurora, will ultimately hit performance and latency barriers. And if you are basing your soon to be successful, rapidly growing e-commerce site on this technology, those limitations could spell slowdowns and lost revenue.

The cloud is the perfect place to offer seamless scale-out performance. Kudos to Amazon for tackling the storage and read aspect of this paradigm. Perhaps version 2 or 3 will start to offer a true seamless scale-out architecture as ClustrixDB does today. Unlike Amazon Aurora, ClustrixDB was designed from the start to scale-out storage capacity, throughput, and overall performance linearly to accommodate an unlimited number of users, transactions and data thanks to the power of massively parallel processing. With its shared-nothing architecture, extra servers can be quickly and seamlessly added in the cloud to accommodate demand spikes while concurrently ensuring a seamless experience for the end user.