[email protected] said:
> Section 3.2. talks about servers that, due to server error or intentional
> malfeasance, sends an error value. The signing does not prevent that, and
> the only way to detect it is by comparing responses from multiple servers
> (as described in Section 3.2). But, if you use a single server (Section
> 3.1), the signing won't detect a server error or intentional malfeasance,
> will it?  

Short answer:
  Yes

Long answer:
  How paranoid are you?

I'm not an expert in this area.

There is something about time that makes verifying answers complicated.  
Banks have audit trails.  Math proofs can be checked by other 
mathematicians.  I don't see how to check timestamps without knowing the 
time.

If you use 3 servers and 2 outvote the 3rd, you are still trusting those 2.

I think that will catch most non-malicious errors.

In the US, the NSA has arm twisted phone companies to install taps on 
internet trunk lines and keep quiet about it.  You don't have to be all 
that paranoid to think that they could arm twist the people running 
Roughtime servers to run patched versions of their code.  That might be a 
key step in catching a spy or a pesky journalist or whistleblower.

Ideally, you would use 3 servers run by 3 different organizations using 3 
different software implementations running on 3 different operating 
systems in 3 different well guarded buildings in 3 different countries...


-- 
These are my opinions.  I hate spam.



_______________________________________________
Gen-art mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to