add parameters

[Imported from Trac: page OneHundredYearCryptography, version 7]
zooko 2010-08-10 16:44:23 +00:00
parent d5b61ffaf9
commit 929a119bb3

@ -29,7 +29,24 @@ David-Sarah has proposed to use hash-based digital signatures.
* <http://tahoe-lafs.org/pipermail/tahoe-dev/2010-June/004509.html> * <http://tahoe-lafs.org/pipermail/tahoe-dev/2010-June/004509.html>
* <http://tahoe-lafs.org/pipermail/tahoe-dev/2010-July/004587.html> * <http://tahoe-lafs.org/pipermail/tahoe-dev/2010-July/004587.html>
Brian and David-Sarah wrote [a simulator](OneHundredYearCryptography)/hashbasedsig.py or two to explore performance trade-offs in hash-based signature parameters. The output of one run is this: Brian and David-Sarah wrote [a simulator](OneHundredYearCryptography)/hashbasedsig.py or two to explore performance trade-offs in hash-based signature parameters. The output of one run with the following parameters is this:
```
# range of hash output lengths
range_L_hash = [128]
lg_M = 53 # lg(required number of signatures before losing security)
limit_bytes = 480000 # limit on signature length
limit_cost = 500 # limit on Mcycles_Sig + weight_ver*Mcycles_ver
weight_ver = 1 # how important verification cost is relative to signature cost
# (note: setting this too high will just exclude useful candidates)
L_block = 512 # bitlength of hash input blocks
L_pad = 64 # bitlength of hash padding overhead (for M-D hashes)
L_label = 80 # bitlength of hash position label
L_prf = 256 # bitlength of hash output when used as a PRF
cycles_per_byte = 15.8 # cost of hash
```
``` ```
w A h T lg_N L_GMSS (bytes) c_S_mean (Mcycles) c_S_worst (Mcycles) score w A h T lg_N L_GMSS (bytes) c_S_mean (Mcycles) c_S_worst (Mcycles) score
13 18 1 62 258.5 97524 (12191) 97272 ( 98.36) 193095 ( 195.26) +5.4 % 13 18 1 62 258.5 97524 (12191) 97272 ( 98.36) 193095 ( 195.26) +5.4 %