I would like to try and clarify the current situation with the "Berlin
A5/1 Rainbow Tables", and the data formats that are being used for
exchange. There are a few steps in generating the tables. The first step
of generation produces unsorted "multifiles" - all output is coarsely
sorted into 256 buckets at 11 bytes pr chain.

An unsorted table is ~90GB - and after sorting (using the sort2 program)
merges have been removed, and you are left with 66.7GB of multifiles.

For the test lookups, I introduced a slighlty more compact format "SSD"
@49.5GB - which only uses 8 bytes pr chain with an accompanying 97Mbyte
index file. For practical reasons this more compact format was also used
in some initial distribution.

Originally 2 programs were made; SSDwriter & SSDexploder to convert
between SSD and MultiFile formats. But when we started adding more
formats, having a single program for each conversion would not be a good
idea.

More recently an even more compact format had to be created in order to
fit more tables in our 2TB limit. This format is a "raw delta" format,
and does not currently have an index, meaning it can only be used for
transportation. Also with this new format, I made a the conversion tool
"TableConvert" - that is able to do any conversion required.

But our intentions are to use a format derived from "raw delta" for the
actual lookup. Raw delta compresses a chain to 54.085 bits, @41.9Gb
allowing us a total of 47 tables on a 2TB disk, but is not the most
efficient for lookup, in part because chains can span the disk block
boundaries. This means that yet another format (at least) will be added,
which may displace the "raw delta". For this reason "raw delta" should
be considered a transitional format until we have figured out the best
way to create an "indexed delta" format. The SSD format will also most
likely go away when we get the new lookup code working.

To convert from "raw delta" to SSD for test lookups, the following
command can be used:

TableConvert ds 100.dlt /dev/blockdevice index.dat

"ds" are input / output format codes, where legal codes are:
   d - raw delta
   m - multifile
   s - SSD

I have started distribution of tables in the "raw delta" format, and the
MD5 checksums for those files are: 

15768f7f5af4e5467304c7162688425d ./100-172/140.dlt
55b1bed783fd94a1feb06e70af5cd789 ./100-172/148.dlt
cec43159aa1d7ca4982f1c7417b8dd09 ./100-172/156.dlt
9558f5a7152ab8dddc6e5ff819a3fa5f ./100-172/164.dlt
307cc99d1724bd1b460d7f39d6121048 ./100-172/172.dlt
f0dbac482c9f6d6626a70bf8115515b4 ./100-172/132.dlt
9db534d5866cc6c8b18eef0bbf6aceae ./100-172/116.dlt
d89f2f62a471734fc86844261e709719 ./100-172/108.dlt
6d36bca865ad5f5b3b95e50be5555f0a ./100-172/100.dlt
0c93864ad638b531dedf10d85880ee79 ./100-172/124.dlt
0cbd427fb98dc9cec05bc8cd9eec53c5 ./180-260/180.dlt
8771eba46785c636e606560265efcb6f ./180-260/188.dlt
7f4fcaec2da609dd794ecff07f1ad890 ./180-260/196.dlt
69b40dc68cd9d26006f9b1f70ab676ad ./180-260/204.dlt
032414372c864417a502ef1eaaa42454 ./180-260/220.dlt
6afc1ca6c0005bf2df6782371c56700d ./180-260/230.dlt
c6a95cda3e55079295cf97706dc9de2e ./180-260/238.dlt
f54c106d3ea9ef442330c5156c14c832 ./180-260/250.dlt
2cf1377b93139fb2a7ec9f687b272d3c ./180-260/260.dlt
9e9861bd2272b58737629b7c34eb1064 ./180-260/212.dlt
af8167abdfd178a01b2bb3a2ee4f59c8 ./268-364/268.dlt
d422218e33cb6c421108469940c03d41 ./268-364/276.dlt
6995a5445330b78380b3ba16642318e9 ./268-364/292.dlt
1cc82ad35788afc4acab95f481e3de9d ./268-364/340.dlt
183a3427843105942e9b0ab0a852c7db ./268-364/348.dlt
771ec801ac043291b9f134f75f82a31d ./268-364/356.dlt
59913ab26509afaecd8861a6877b06c6 ./268-364/364.dlt
75340dda19dc1b1ada8108d47461562a ./268-364/332.dlt
2ecc836a993a3f71cd710c4455e9780e ./268-364/324.dlt
5bc5814694b412cfb4ba52198492ba1b ./372-500/404.dlt
44f06b97410bda5b029754de9cc87db4 ./372-500/428.dlt
435e2d2755aa5ce9357889bef5ee534f ./372-500/380.dlt
98de5b58c98fafc5ada3897eddf7f31a ./372-500/396.dlt
4518baa2a1156451cbc0952eb216cb5f ./372-500/412.dlt
1bbceb9afda870b76b7e9928409183de ./372-500/372.dlt
3985446a74105d15f6c4b8bd0467771b ./372-500/420.dlt
fabeaa1792045a0ab5471aeac7271481 ./372-500/388.dlt
a0006b7b58e01b97d6bce1b43b015343 ./372-500/492.dlt
fa47f2242230b69c5ae7a1196ea0311f ./372-500/500.dlt
c3c48ba123a1ebd9ea6f0abaa56ac475 ./extra/436.dlt 

-- Frank

_______________________________________________
A51 mailing list
[email protected]
http://lists.lists.reflextor.com/cgi-bin/mailman/listinfo/a51

Reply via email to