Tuesday, May 27, 2014

NSA funds CSEC big-data research?

We know that NSA provides funds to CSEC from time to time, but the purpose of that funding has not been revealed.

The amount of money provided by NSA is tiny in comparison to CSEC's overall budget, but it still adds up to a significant sum. As discussed here, in fiscal year 2012 NSA gave CSEC approximately $350,000. Since 2002, CSEC has apparently received more than $11 million from the “Five Eyes partnership”, most if not all of it almost certainly from the NSA.

According to one NSA document, NSA "at times pays R&D and technology costs on shared projects with CSEC". The Canadian government similarly describes the Five Eyes money it receives as "investments received from partners for cryptologic research and development".

However, these very general descriptions don’t tell us very much about what the two agencies are actually working on. The term "cryptologic" could refer to anything in the Canadian Cryptologic Program, which covers everything that CSEC does.

The CSEC “airport wi-fi” document revealed in January (IP Profiling Analytics & Mission Impacts), on the other hand, may provide an indication of what NSA has recently been funding.

As the image above shows, the "IP Profiling Analytics & Mission Impacts" document reported that NSA provided “launch assist” for the Collaborative Analytics Research Environment (CARE), “a big-data system being trialled at CSEC”.

Launch assistance could mean a number of different things, including shared research efforts or the contribution of hardware and/or software, but provision of funding for the project is certainly a possibility. The CSEC document was written in the middle of U.S. fiscal year 2012, so the timing of the $350,000 transfer seems to be about right.

The goal of the CARE project seems to be to develop improved methods of processing and analyzing huge sets of data without overwhelming the available computer resources. The "game-changing" results reported for the research suggest that methods for handling big data are rapidly improving, making it feasible to conduct kinds of analysis that were computationally infeasible in the past.

I would welcome comments from any readers who can help to interpret the description of the CARE system provided in this slide.


Post a Comment

<< Home