[CQ-Contest] authentication for log submission

[Yannick DEVOS (XV4Y)]

Sorry Dick, that was you I wanted to quote about LotW.
As you state, forging a fake log that seems genuine enough but harmful to the score is something with a low probability.
Making the submission delay shorter for the top-contenders is also a way to reduce this probability.

73,
Yan.
---
Yannick DEVOS - XV4Y
http://xv4y.radioclub.asia/
http://varc.radioclub.asia/
> Don, you are a man after my own heart! I'm totally paranoid about security
> scenarios, which is one reason LoTW security is so tight.
> 
> However, I think the scenario you describe isn't plausible. As I understand
> it, the scheme would be used in a case where two stations are close in
> score. It would involve the second-place station uploading an altered log
> for the first-place station after the first-place station uploaded his/her
> real log. The altered log would be mostly the same as the real log, but with
> a few QSOs omitted or a few calls or exchanges busted. It has to be an
> altered copy of the first-place station's log because if the log was a
> complete fake, none of the QSOs would match during log checking. I'm sure in
> such a scenario, where the log for a high claimed score generated a huge UBN
> deduction, the log would be visually inspected by a person. That would alert
> the contest sponsor and the first place station that something very fishy
> had happened. Therefore, the altered log must be a copy of the first-place
> station's log, with just enough QSOs altered to affect the standings.
> 
> The question is, how does the second-place station get a copy of the
> first-place station's log? ARRL doesn't publish logs. I think CQ WW waits to
> do that until after log checking, but I'm not sure. If they don't, they
> should.
> 
> The only way I can see pulling this off is if the cheater had a confederate
> who operated in the contest and generated a relatively high score, slightly
> less than the second place station's score. The confederate's log would be
> submitted under the call of the first-place station. But that depends on how
> the log-checking software determines the call sign. Is it solely from the
> Cabrillo header, or is the call in the header checked against the sender's
> call in the individual QSO records? In any case, this is a very risky ploy
> for the confederate because if the log is inspected visually by a person,
> which I believe happens with most high-scoring logs, the jig will be up.
> 
> So, I don't think it would work. However, if there is a scenario that would
> work, the best solution would be to use the LoTW authentication system for
> log submissions. That would have the dual advantage of verifying the call
> sign in the log and allowing automatic submission of the log to the LoTW QSL
> system (something many have asked for.)
> 
> 73, Dick WC1M

Le 6 juin 2012 à 18:18, Yannick DEVOS (XV4Y) a écrit :

> Dear Katsuhiro, Michael,
> 
> Katsuhiro, you are right this is a serious security flaw in the way the log submission are handled.
> It can lead to spoofing (someone use your identity to upload logs) and flooding (trying to overload the server).
> However, as Michael stated, this issue is mitigated by the difficulty in forging logs that could be really harmful to the whole contest integrity.
> A well designed server will also discard bad crafted logs without too much database load.
> 
> The only way to have a 100% secure system is the way LotW goes.
> However it is not easy to handle and will increase contest handling costs a lot.
> 
> If I were a contest server administrator, what I will do is the following :
> - for 95% of the participant, nothing at all just like today
> - optionally, participant who want to secure their log can request an "ID token" upon sending one hand written dated and signed scan of their license
> - an additional filed in the Carbrillo format will content this token and it will be checked while the log is processed
> 
> This is not fully secured as someone can "sniff" the token on the network (it is never crypted in the process) or hack the contestant computer and copy it.
> However if someone is serious enough to do this, this means all the security on the server and the contestant computer has to be checked, and this raises the bar significantly.
> For me, it add a fair level of authentication for a marginal managing cost increase.
> 
> 73,
> Yan.
> ---
> Yannick DEVOS - XV4Y
> http://xv4y.radioclub.asia/
> http://varc.radioclub.asia/
> 
>> I'm not sure this was discussed before, but this reminds me that
>> someone who has malicious intention may submit other station's
>> log to defeat the station after first submission by actual station.
>> There looks no authentication method to verify the station for major
>> contests(please correct me if I am wrong).  Complicated method to
>> authenticate the station may lead decreasing the number of log
>> submission, so this may not be applied to all stations.  But I think
>> there should be some method to authenticate at least for stations who
>> want to win a prize.
>> 
>> Please ignore this message if my concern is baseless fear, the
>> contest sponsors have already taken care of this, or we can trust
>> everybody since we all have good morals.