Summaries - Research
Back Sector Discrimination Seedling
|Division||Graduate School of Operational & Information Sciences|
|Investigator(s)||Garfinkel, Simson L.|
|Sponsor||Defense Advanced Research Projects Agency (DoD)|
This proposal is to develop and test a principled approach for identifying individual disk sectors. Some sectors (such as the sector of all NULs) commonly occur in many files; other sectors (such as sector encrypted data) are "unique"-probabilistically they will never occur elsewhere on the planet unless a bit-for-bit copy of that sector is made. Still other sectors (like a sector from a Wikipedia article) may be unique themselves, but may contain internal structure that allow them to be identified with a high degree of accuracy. This project would attack the Sector Discrimination Problem with three different approaches, build a prototype sector discrimination engine, and test that engine with two datasets. My testsets. My testable hypothesis is that we can dramatically improve the speed, accuracy and coverage of sector discrimination from the current method, which relies on the identification of file headers and unique hashes from the NSRL data set.
This project will use DARPA funds for computing sector hashes for all of the NIST NSRL, improving the Bloom Filter technology that I have been working on, and exploring a new algorithm that Vassil Roussev at Tulane has developed for identifying high-entropy features on the sub-sector level.
|Publications||Publications, theses (not shown) and data repositories will be added to the portal record when information is available in FAIRS and brought back to the portal|
|Data||Publications, theses (not shown) and data repositories will be added to the portal record when information is available in FAIRS and brought back to the portal|