Richard Tibbetts, CTO and co-founder, StreamBase Systems

Many customer interactions today are completely electronic, even high value interactions, reducing the opportunity for organizations to build personal relationships and provide high service. A key example of this is equities trading, where the role of sales traders is often to respond to problems and failures reported by customers. In order to build relationships in these situations, customer service representatives require complete real-time analysis of the customer's experience and actions, and every other system that could impact the customer experience, potentially dozens of different data sources.

A View Server is an architecture pattern for using real-time analytics to create a set of materialized, pre-integrated, pre-analzyed views of data relevant to end users like sales traders, operations staff, and others. Hundreds of gigabytes of data can be stored memory resident across a cluster of machines, maintained and updated with sub-second latency. Using a desktop or web interface, nontechnical users can easily query, understand, and react to real-time data, seeing not only the results of their query, but watching live updates to those results as problems are resolved.

Imagine having a customer on the phone, while the customer is also on a website or mobile device, and immediately seeing the results your actions, the customers actions, and other systems that may resolve the problem. The View Server can also be used to drive alerting and case management, so staff can resolve problems before customers notice them and practice management by exception. The StreamBase LiveView View Server has been deployed in support of execution services at a major international brokerage firm, handling tens of thousands of events per second and millions of customer interactions per day.