Scaling Your Application with Riak

Operations fanatics at heart, we built Riak to ensure that scaling not just up but down would be as close to trivial as possible.

Typical databases do not truly scale in a linear way. This stems largely from a skewed understanding of what it actually means to scale. For example, simply adding more machines to handle a massive number of requests does not equate to scaling up if you are sacrificing throughput. Additionally, when first building your application, it is almost impossible to predict the number of users you will be have in one month, six months, one year, etc. So the best way to plan for scaling problems is to use technology that delivers predictable performance.

That is why Riak features pluggable scalability. What this means is that in order to deal with heavy load to your site or application, operations personnel need only to enter a shared secret into a racked server and Riak will take care of the rest. In a Riak cluster, if you double the number of servers, you double the your amount of storage while maintaining the same throughput your customers and users have come to expect. Scaling with Riak is linear, predictable, and cost effective.

And, because Riak's data schema is flexible, you no longer need to change your data design to add new data formats. With Riak, it's a simple PUT operation over HTTP to add new data types. Reducing operational complexity also results in reducing the costs related to maintaining your database.