I would like to give a short update on what goes on a little behind the
scene. Eagle eyed participants may have noticed some clues in the SVN
repository and source tree.

Firstly a lot of though and discussions have taken place around MvdS's
post to the list about applying extra clockings in the rainbow table
round function. And a consensus has more or less been reached, giving
merit to the ideas as posted to the list. And I would personally like to
thank MvdS for his theoretical contributions to the project.

It has however taken quite a bit of time to verify these findings, and
tune the table parameters for optimal performance. Karsten has been
toiling over this part, and is probably best suited to explain the
choices and candidates. Sascha and I have computed various test tables
to verify some of the assumptions. Currently I am computing my first
"production table" - and will then have to do some further verifications
on the results.

At this point I should underline, that the current generation of tables
are still good, and will be used for lookup. But the ATI tables are
slightly different, and will need some special attention at the sorting
and lookup stage. I have 10 completed tables, and it would probably be
best if the other ATI tables be sent my way after sorting to be
integrated in the lookup process.

However chances are that we will do a transition to tables with a new
format. I hope this will be done in an orderly manner, even though it is
hard to give a time line for how and when this will happen. I have maybe
even gone out on a limb here in announcing this somewhat prematurely.
But I think we have a duty to deliver timely information, as well as
managing expectations.

The current proposal entails computing fewer but larger tables, so for
the distributed project the workload will have to be sliced vertically
as well as horizontally, meaning a fair amount of work may have to be
done in that area. So please don't ask when the "new client" will be
ready.

The good news is, granted that the new table format delivers as
expected, we should be able to achieve the desired total coverage at an
accelerated rate. Even so, I would like to remind anyone eager to start
computing "new tables", pushing forward the time of reaching desired
coverage by a few months, means very little if the intercept capability
required for using those tables is still lacking....

regards,

  Frank




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

Reply via email to