I understand that Intel TXT provides a Dynamic Root of Trust Measurement (DRTM). Does the Linux integrity measurement architecture do logically similar things? If not, what are the differences?
No, they are complementary.
DRTM provides a Root of Trust on which you can build applications/use-cases such as Integrity measurement of your entire linux system (IMA). However, it is not part of it.
I think the simplest way to explain this is with the following:
- IMA needs the kernel to evaluate the integrity of the system (e.g. userland);
- DRTM (or SRTM for this matter) can validate the integrity of the kernel itself. See tboot.
I suggest reading this to have a feel for it - the "Value Proposition for Linux" more precisely.