Re: Why won't people use clouds: The grey part of the silver lining

Oh that is certainly not true! While it is possible that all the
systems will be equally compromised, it isn't the only possibility. I
can just as well build a single system out of thousands that will be hacked.

Chuck

Sassa NF wrote:
> Define a metric for "secure".
>
> More opportunities to be compromised? Maybe. But better chances of
> success?.. That may be true only for non-homogenous systems, and on
> certain conditions. In a *homogenous* system every component perhaps
> has the same probability of successful compromise. Having N replicas
> of the database with exactly the same probability P of compromise does
> not change the chances of compromise (N*(1/N)*P, which is "N" times
> the probability that the particular replica was chosen multiplied by
> the probability of success). It is fair to assume that in a homogenous
> system the probability of attack is distributed uniformly (1/N).
>
> For the messages in transit to N destinations to be significantly
> easier to compromise than one, N must be really large, if we were to
> consider analysis of multiple encryptions of the same message.
>
>
> Sassa