Abstract

Morgan Stanley spend around $3 billion each year on technology for their global securities trading business. The majority of the application development takes place in-house, using bespoke development frameworks that focus on performance, stability and the ability to deliver new functionality very quickly. In this talk representatives from Morgan Stanley will present practical example problems for discussion from their existing systems. They will also talk about some of the more challenging distributed systems problems they face when designing the large scale trading systems of the future.

In the second part of the talk we will describe in more detail the architecture and cross-language implementation of the real-time high volume data system presented in the first talk, with emphasis on the pitfalls and lessons learned while implementing the system. We will also highlight potential research problems we have encountered in the course of the project.