Talk: Patterns for scalability and availability in (trading) systems

This talk will walk through a number of useful nontraditional approaches and architectural patterns to building systems that combine high availability, high throughput and low latency.

Throughput and low latency is often contradictory and some of the tradeoffs you do usually comes at expense of either one or the other. This presentation covers approaches used in large scale trading systems where there are requirements to process hundreds of thousands of business events per second and reliably react to them whilst capturing and reacting to thousands of business transactions per second while maintain and distributing data to tens of thousands of concurrent users in an environment where milliseconds matter.