The Verdict Is In: Database Sharding Is Complex, Expensive and Risky
Database sharding has been rumored to be a highly scalable approach for improving the throughput and overall performance of high-transaction, large database-centric applications. But what enterprises have learned over time is that sharding introduces significant opportunity costs and risks by requiring additional application code and operational complexity.
Scaling Databases and the Sharding Deep Freeze
Most internet infrastructure projects start with the idea of connectedness. This includes features like ownership, sharing, notifications and updates, privacy, and community. There is a database layer (usually MySQL) that holds all of the structured information necessary for the application to represent the above functions. Software architects can start down a sharding path and pretty soon realize they’re stuck in a “deep freeze.” From then on, development to deployment takes longer and is more involved. Once a sharding backend has been introduced, the effort required to get the most basic features developed continues to go up and up.
Download our white paper to learn more, Scaling Databases and the Deep Sharding Freeze.
Learn more download our white paper, A New Approach: Sierra Clustered Database Engine.
Clustrix Reduces Cost and Complexity by 90% Compared to Sharding
Clustrix has created a distributed database that grows online through clustering. The database itself lives across the nodes in the cluster, allowing the system to not be constrained by the size of any single piece of hardware. This also allows the system to route around faults and be easily managed as a unit.
A Radically Simpler Approach Instead of Sharding
Dramatically reduce the cost and complexity of scaling applications by eliminating the expense and effort of database sharding. Now developers can stop working around the limitations of the database to focus 100% on innovation, and DBAs can manage more DBs and shift focus to the data.