Presentation: Tweet"Banking Case study: Scaling with low latency using NewSQL"
Most assume that horizontal scalability requires application managed sharded databases, use NoSQL databases or using in-memory data grids. But, often this mandates application re-architecture and injects objects or turning access in key-value based access can be quite challenging. This presentation walks through how a bank used a horizontally scalable, memory oriented SQL database (VMWare SQLFire), did minimal changes to its application, injected the Database between the app and its original database and achieved elastic (on-demand) scaling with very low latencies.
We will spend time walking through the design patterns in the application and then go a little deeper into SQLFire concepts - horizontal partitioning policies, how to achieve scalable transactions with colocation, distributing application logic to distribtued data and parallelizing execution, query processing, etc. A lot of concepts are similar to several ""next generation"" SQL products.