Hi, On Thu, 6 Jun 2013 14:05:48 -0700, Matthias Vallentin <[email protected]> wrote: > When constructing a Bloom filter, users specify a desired > false-positive rate and a capacity (max number of elements which > guarantee the FP rate), which the implementation uses to allocate > internal storage. Imprudent parametrization can lead to values beyond > the machines memory limits, which would immediately crash Bro. Should > we expect that users only do reasonable parametrizations by clearly > documenting the effects of the parameters? Or should we create hard > caps in the implementation to avoid users shooting themselves in the > foot?
If you do decide to have hard caps, could those caps be easily configurable? An amount that might crash on one machine today could be entirely reasonable on another in a few years (or even on a more expensive machine today). -- David Eric Mandelberg / dseomn http://david.mandelberg.org/ _______________________________________________ bro-dev mailing list [email protected] http://mailman.icsi.berkeley.edu/mailman/listinfo/bro-dev
