Last year, Nimble Storage announced SmartStack™ for Oracle with Cisco at OpenWorld 2013. Since then, Nimble has been expanding its portfolio of application proof points. As part of that effort, we tested MySQL database on our arrays.
MySQL is the world’s most popular open-source database. It’s being used for ecommerce, online transaction processing (OLTP), and embedded database applications. It is a fully integrated, transaction-safe, and ACID (Atomicity, Consistency, Isolation, Durability) compliant with full commit, rollback, crash recovery and row-level locking capabilities. Many of the world’s most trafficked websites, like Facebook, Google, Ticketmaster, and eBay, rely on MySQL.
Unfortunately, provisioning storage for MySQL can be a time suck. Database and storage administrators spend countless hours configuring storage to get the level of performance – without the usual hit to system availability – that MySQL demands. That means they have to take on the time-consuming chore of examining RAID levels to guarantee there is no single point of failure in the configuration.
That extra work adds to MySQL’s total cost of ownership (TCO). And, there are other factors that contribute to its relatively high TCO. Traditionally, to sustain a high number of IOPS (with low latency) for an OLTP workload required adding several disk drives. The higher the IOPS, the higher the number of drives needed. Many storage administrators have continued the unfortunate practice of creating dedicated silos for data protection, which also adds to MySQL’s TCO.
Nimble has published a best practice guide that explains how our arrays lower the TCO of MySQL Community Server with an InnoDB Storage Engine implementation running on Oracle Linux 6 and the Red Hat 6 operating systems while at the same time delivering the performance they demand. Although MySQL Cluster is not covered, this detailed guide provides a basic understanding of the internal I/O processes that MySQL Community Server uses to store data on disk. It describes how Nimble solutions deliver:
- Adaptive Flash for high read/write performance;
- Inline compression, zero-copy cloning, and thin provisioning;
- Scale-to-fit sizing with non-disruptive upgrades of capacity and/or compute; and
- Instant backups and restores with no-overhead snapshots, and no more backup windows!
Database and storage administrators can use it to provision storage in MySQL environments for optimal performance (with no backup windows). In the next few weeks, we’ll follow up with blogs detailing Nimble’s value-add in database implementations.
- Ibby Rahmani