Post #3949554

2026-07-20 03:45 UTC

@muddle@infosec.exchange Some quick searching will reveal https://www.freerainbowtables.com/rti2formatspec.pdf "Introduction RTI2 is a variable length bit packed rainbow table format developed by www.freerainbowtables.com. It reduces disk usage through the variable length data chains, though each data file has one consistent data chainrow length, and through the use of prefix indexes. It is in some ways similar to the RTI format but designed to be easier to modify items in subsequent minor versions and this flexibility adds a fair amount of complexity. Additionally, it is designed so that a transition away from the file naming for information may be possible as this will be necessary for more complex table sets in the future"

Replies (3)

@muddle@infosec.exchange 2026-07-20 03:53

@thedarktangent@defcon.social OK, thanks. I didn't read the whole thing but I assume that the whole point is that the mapping is surjective? That is to say, if we mapped all possible inputs to all possible outputs, the set of outputs would be at least potentially smaller? I'm looking for a rationale as to why the data set is compressible at all and this seems reasonable at least. Damn. I used to know this stuff.

Open ##3949551
@muddle@infosec.exchange 2026-07-20 03:56

@thedarktangent@defcon.social ... and chain length is proportional to a number of rounds, each of which is capable of being captured into a smaller set of pre-computed orbits? Sorry if I'm forgetting this, but I thought that most salted, hashed passwords only involved a single round? (aight, sorry... back to the books for me)

Open ##3949638
@muddle@infosec.exchange 2026-07-20 04:09

@thedarktangent@defcon.social OK. I get it now. Birthday Paradox (hash collision) implies ever-decreasing circles (aka limit cycles, etc.). So compression is possible.

Open ##3949910