![](bg2.png)
lates most of the terms of services which usually state
like that “each subscriber agrees that not to allow oth-
ers to use her/his member name, password and/or ac-
count.” Thus, the act could involve defeating a com-
bination of technical and contractual access control
measures (Mulligan et al., 2003).
A subtle analysis of potential legal risks for users
who attempt to share account can be found in (Mul-
ligan et al., 2003), which states that the structure of
DRM applications may drive users seeking to engage
in customary personal uses of copyrighted works to-
ward legally questionable behavior. Unfortunately,
there are also some difficulties for the providers to
make a plausible claim against such account mis-
uses, such as how to detect the difference in physical
identities, and to meet some relatively high damage
threshold (Mulligan et al., 2003). This is the reason
that in practice, with regard to non-serious cases, the
providers usually prefer contractually reserved self-
help measures (e.g. terminating a suspicious account)
to law.
Therefore, the point here is how to encourage the
consumer to keep the secret (account and password)
while not to share it among friends, since current
technical solutions and legal measures are difficult to
solve this problem perfectly as we discussed above.
In the work described here, we propose an incentives-
based secret protection system (IBSPS) to solve this
problem. IBSPS can novelly transform the burden of
secrets protection of content providers into the bene-
fit of consumers, and thus stimulate them to keep the
secrets private in their own interests.
IBSPS is suited for applications where a secret
must be protected for only a limited period of time. A
typical example is the accounts and passwords of on-
line subscription services which usually have a period
of validity (i.e., monthly or yearly subscription). Con-
tent files purchased from online shops are not con-
sidered in current IBSPS, because generally speaking,
these files once purchased, will be valid for ever.
It should be noted that we do not want to use IBSPS
to replace current terms-of-use policies, or other legal,
technical protection measures. Rather, we aim to add
IBSPS as an additional layer of protection for DRM
systems that already have such safeguards.
1.1 Related Works
Our work was primarily inspired by SPIES (Margolin
et al., 2004), which aims to provide an economic neg-
ative incentive to not share the secrets such as pass-
words. The main idea of SPIES is to require the con-
sumer to place an additional security deposit into a
trusted escrow account beforehand, and then if she
shares the secret with other unauthorized users, she
will totally lost some money in return. The mecha-
nism is that, every one who has a copy of the secret
(including the unauthorized users who got the shared
copies from the authorized consumer and did not put
corresponding deposits into the escrow account) can
register to the escrow service and receive a share of
the deposit after the secret protection period. A de-
posit payment function is proposed to guarantee that
any kind of collisions or multiple registrations will re-
sult in totally losing money as for the authorized con-
sumer.
The deposit idea of SPIES is derived from some tra-
ditional existing applications, such as entertainment
reviews before the product is available to the general
public. However, when it is applied to the password
management of online subscription services, it has
several limitations as follows.
First of all, the additional security deposit itself is
a big disincentive for the user to choose such service.
In SPIES, the deposit v is set to v = c(n +1), where
c is the price of the service, and n(n>2) is an es-
timated number of users sharing an account such that
it will easily result in concurrent usage of the account
detected by the provider. Correspondingly, SPIES as-
sumes that the provider will not technically prevent
simultaneously access to an account beforehand
1
,but
to allow and detect such concurrent usage in order to
deactivate the account afterwards, which is also an
important mechanism of SPIES to guarantee that an
authorized user can never sell her account to more
than n other peoples so as to make profit. Some trou-
bles may occur in this case, for instance, how about an
honest consumer forgot to log out her laptop in home,
and then go to her office to open another desktop and
log in again with the same account?
Secondly, SPIES shifts some potential unex-
pectable risks of secret (password) disclosures to the
users, such as new virus, OS or software security
bugs, and even inside leaks of the provider. In any
above case, the user will risk an additional big loss of
her deposit, which usually exceeds the normal price
of the service itself. This punishment is too strict
and may bring about some legal troubles to distin-
guish the responsibilities. As a remedy, in (Margolin
et al., 2004), it suggests that the provider could give
the consumer one day or more to report a stolen pass-
word. However, it is still too strict to require each nor-
mal user (not computer expert) has an ability to detect
such security intrusions in time. In contrast, a more
reasonable case may be that, even the password has
been stolen due to the user’s personal improper pro-
tections (e.g., not updating security packages in time),
the user may accept to lose at most the cost of the ser-
vice itself (i.e., could not enjoy the service anymore),
1
In contrast, most current online subscription services
just simply kick the first user out automatically if a second
user logs in using the same account, or prohibit repeated log
in if a user is using the account.
WEBIST 2006 - SOCIETY, E-BUSINESS AND E-GOVERNMENT
6