[Dshield] Crypto Question

David Brodbeck brodbd at u.washington.edu
Fri Mar 6 17:31:02 GMT 2009


On Mar 5, 2009, at 6:52 PM, Valdis.Kletnieks at vt.edu wrote:

> On Thu, 05 Mar 2009 10:19:52 PST, David Brodbeck said:
>
>> To me this seems like basic "defense in depth."  Engineers would call
>> it redundancy.
>
> Which is why your programs all implement two separate 'sort' routines,
> invoke them both, and compare the results, just in case one screws  
> up, right?
>
> Oh, you don't need to do that, because you trust the first sort  
> routine,
> and if it was buggy you'd just replace it, because carrying around 2  
> and doing
> it twice is just a pain in the ass for no real gain? Hmm. Gotcha.

This presumes that it's fast and trivial to swap out one algorithm for  
another when a problem is found.  That's not always true.  Having two  
hashes buys you time in the event one of them is broken.

Why does your car have three braking systems (two hydraulic, one  
mechanical)?  If one was broken, you'd fix it, therefore you should  
only need one, right?  After all, having three is a pain in the ass  
for no real gain...

-- 

David Brodbeck
System Administrator, Linguistics
University of Washington






More information about the Dshield mailing list