Although fraud is a large, complicated, and application-dependent subject, everyone agrees that it is better to avoid heavy financial losses, unhappy customers and churn. That goal is well-served by performing fraud detection continuously and as near to real-time as possible.

SQLstream provides an ideal platform for continuous and real-time fraud detection solutions:

• Continuous execution of complex, time-sensitive matching rules, such as detecting someone apparently using a credit card or mobile phone in two locations at the same time.

• Capturing and replaying data, with configurable levels of detection targeted only at messages that, for example, trigger some potentially coarse-grained, potentially-fraudulent filter.

• Distributed, optimized, load-balanced processing.

• High data volume throughput.

• Easy management and transformation of multiple formats of data, enabling uniform processing of diverse sources of streaming data

• Easy management, maintenance, and enhancement of streaming queries and distributed processing nodes.

• Easy extensibility and interfacing with specifiable external programs for case-specific detection and analysis

• Easy integration with heterogeneous data sources, existing (or new) streams and external databases, log-files etc.

• Simple integration with alerting and real-time action-related systems, existing or to-be-developed