Microprocessors (CPUs) have Single Instruction, Multiple Data (SIMD) capabilities that are typically used for audio and video processing. We have developed software and patented techniques to enable SIMD processing for text and variable length data. Processing can now be done 256 characters at a time rather than one character or word at a time, resulting in dramatic improvements (up to a six-fold improvement in throughput) to widely used ETL database operations (data validation, parsing and format conversion) and in other big data applications. With our software running on servers, desktop machines and even handheld devices, text format conversion can be dramatically accelerated, pattern searching and regular expression processing can be performed much more rapidly, while also obtaining a dramatic decrease in energy consumption. See Acceleration of ETL Processing Software for more details.

Our open-source software and patented technologies (US Patents 7400271, 7728738, 7783862, 7898441, 8077061; Chinese Patent ZL200680030295.0) are available for deployment by end-users and distribution by software developers on standard open-source terms through our patent covenant. The transparency of functionality of open-source software (there are no "black boxes") can be very important to verify that the software meets all customer needs, particularly when it comes to low level routines. We work with partners on commercial licensing opportunities for combining this software with hardware (including mobile devices) or proprietary software.

Visit our Products page to find out more about our accelerated version of XMLWF for XML wellformedness checking, about icXML, a very fast XML parser that does not need any specialized hardware, and about icGREP, our accelerated version of the familiar GREP utility program used for regular expression processing. Additional information about how our technology works can be found on our Technology page.