When Milliseconds Count, Multiple Masters Ensure High Database Availability
If your website is down for “tens of seconds,” or even a few minutes, what does that mean for your business? While it can be an inconvenience for your existing customers, it can also result in lost new business. What happens to your Google AdWords or organic SEO conversions when people click on your link but can’t access your website at that moment? Potential customers who visit your website and get an error message will likely move on to the next company instead of coming back to your website at a later time. How many customers are you willing to lose? For most companies, the answer is none.
That’s why for mission critical MySQL-based applications it is essential to invest in the right database solution that ensures truehigh availability.
Some cloud services claim that they offer high availability, but when you look under the hood you’ll find just a single database master using locally redundant storage (LRS). With this minimal approach you do at least get three copies of your data maintained on three different storage devices, which protects against any single drive failure. When the database master fails however, these services spin up a new one and attach it to your storage, a process which can take “tens of seconds” to complete. Whether that means 30 seconds or several minutes is anyone’s guess.
The key to keeping your database online in dynamic cloud environments like Microsoft’s Azure or Amazon’s AWS is the powerful combination of redundant cluster topologies plus sub-second failover routing.
With ClearDB’s Community and Production tiers, we deploy multiple database masters, each connected to its own LRS storage. With a two-master cluster, that means you have six different copies of your data. These masters are also geo separated by default and stay in sync with each other, so if there is a regional outage in one area, your database stays online and available in the other.
Our proprietary CDBR technology (ClearDB Router) is a smart traffic manager that monitors the health of master nodes in a ClearDB database cluster and automatically re-routes traffic in the event a master becomes unavailable. This automatic failover process typically occurs in just milliseconds, ensuring that an application can continue to run uninterrupted even if its primary database master should completely fail.
This technology, combined with multi-master and geo-distributed cluster topologies, make 100% up-time in the cloud possible with ClearDB’s popular MySQL-as-a-Service. That’s why ClearDB’s Production tier clusters are backed by a 100% up-time SLA.
With mission critical applications that can’t afford “tens of seconds” of downtime, it’s important to have database redundancy across geographical regions to protect against service outages, network failures and even natural disasters. Customers who subscribe to a database service want a fully managed experience in which their database stays online, no matter what. This means that anyone offering such a service must be investing in the resources and system tools to proactively manage and balance cost and performance with dynamic application resource requirements.