On Wed, Jan 11, 2012 at 2:50 AM, Olaf TNSB <[email protected]>wrote:
> I *think* I want to be able to define both physical proximity and network > cost. That is, I want to be able to specify that enough shares are on my > local nodes for me to download files (cheap bandwidth) BUT that there are > enough shares elsewhere (i.e. far enough away) that a local disaster > doesn't destroy all copies. > I didn't get into modeling network bandwidth, but my loss modeling paper does provide a mechanism for calculating the probability of loss of a file whose shares are laid out in any conceivable way -- well, to the degree that you can accurately define and measure the risks. Given a configuration language for defining node relationships and failure probabilities, it would be possible to compute the estimated reliability for any given share placement, and then different placements could be compares to determine which ones meet the desired reliability threshold. -- Shawn
_______________________________________________ tahoe-dev mailing list [email protected] http://tahoe-lafs.org/cgi-bin/mailman/listinfo/tahoe-dev
