Blazing-fast stream processing to revolutionize real-time data analytics.
PanJoin is a new method for joining data streams more efficiently. The goal was to improve the speed and accuracy of combining data in a stream. PanJoin uses specialized data structures to reduce the amount of work needed during the process. It supports different types of data matching, not just equal matches. By implementing a hardware-friendly structure called BI-Sort on FPGA, PanJoin outperformed other methods by a huge margin. In testing, PanJoin was over 1000 times faster than previously proposed stream joining techniques, and it dealt well with unevenly distributed data.