The single node backend is dying or dead. A single instance of a database is no longer sufficiently available, resilient or elastic to cope with the quantities and changing requirements of data that modern applications demand. A single point of failure is no longer an option.
To cope with this, developers have typically taken the 'traditional' database they began with and layered on other technologies such as document stores and better search capabilities. This solves their requirement problems but creates a brittle and complex series of interconnected parts that may not always work together well and are a nightmare to scale, upgrade and keep data in synchronization.
In this presentation I will look at the current best practice techniques for solving the data persistence at scale issue, such as replication and synchronization. I will then look at emergent technologies that aim to solve the problem in a more graceful and cohesive manner. This will include mentioning our own open source solution as well as other open source options and the manner in which they function.
