Information Technology Reference
In-Depth Information
4 Probabilistic Verification
Message verification only applies to vehicles that are unable to check directly the
received information. That is to say, when a vehicle receives a warning message
about an incident that is not under the coverage of its antenna, and wants to
confirm the authenticity of the received message, it has two options depending on
its location with respect to the source of the warning. First we consider the case
when the receiver cannot confirm the information directly, but it is so close to the
supposed event that it has to make a decision quickly because in a short period of
time it will be too near to avoid it. In this case, if a vehicle receives an aggregated
packet signed by several vehicles, it should use a verification mechanism fast
enough to verify the signatures of the packet. As we already mentioned, on the
one hand it is inecient to verify all the signatures contained in a packet, but
on the other hand it is necessary to verify the information before accepting it
as valid. In order to fix this problem, only a few signatures are proposed to
be verified according to a probabilistic scheme. Secondly, we consider the case
when agents are far enough from the hazard so they behave according to the
store-and-carry paradigm, collecting evidences about the hazard in the form of
aggregated packets. In this case, the vehicle has enough time to collect and check
several aggregated messages before it has to make a decision. In particular, it has
to check some signatures for each received aggregated packet as in the previous
case. However, in this case the vehicle may perform more complete verifications
that will provide a higher level of certainty.
As aforementioned, probabilistic verification will be only used by vehicles that
are unable to verify directly the information that reaches them. The proposed ver-
ification algorithm uses threads, which are lightweight processes that allow a con-
current execution for a faster execution of the whole protocol. In the algorithm
shown below, H [ i ] denotes a thread for the variable i that takes an integer value
between 1 and n ,where n denotes the number of aggregated signatures. When
an agent receives a message and decides to verify its signatures, the main process
launches as many threads as signatures the message contains. Before the main
process launches the threads, it checks whether the message contains enough sig-
natures to determine whether the message has been signed by a significant num-
ber of agents. This minimal number of signatures to validate a message shows the
degree of closeness between data packets referring to the message, and is usually
called intimacy level in trust literature. In our probabilistic verification scheme,
such an intimacy level will be given by a combination of two factors: the average
number of authenticated vehicles and the space-distance between the locations of
the receiving agent and the event announced in the aggregated packet. Thus, in
order to calculate the threshold for the minimal number of required signatures, for
example the agent could compute the average number of authenticated users per
minute in the current session and the SD between both aforementioned locations.
Then, the intimacy level might be given by the average of users when the distance
is below 30 m. For distances between 30 m and 1000 m the level can decrease lin-
early with the distance, until reaching the minimal value of 2, which might be the
intimacy level for any distance greater than 1000 m.
Search WWH ::




Custom Search