User:RebranchingJiiil/sandbox

Reputation Management Systems (RMS) are susceptible to numerous forms of network attacks and it is important to be aware of such problems when attempting to implement a successful RMS.

Re-Entry Problem
Many RMS algorithms are susceptible to the re-entry problem in which an entity may leave the network after obtaining a negative reputation and “re-enter” under a new identity.

“Sudden Exit” Problem
A component of the re-entry problem, some RMS algorithms suffer from the “sudden exit” problem in which an agent misbehaves and then escapes the network without being penalized for its potentially detrimental actions. In such cases, the agent does not rejoin the network and the identifying information of the agent may not be enough to trace down and castigate the attacker.

Multiple Identity Problem
The multiple identity problem stems from the re-entry problem in that an attacker may be abusing the network it is participating in from multiple access points or agents of differing identifications.

Collusion
Collusion is identified as one of the greatest vulnerabilities of any RMS. In collusive attacks, entities in the network agree to work together in order to obtain higher reputations, boost the reputation of a peer, or lower the reputation of another entity in the network. This goes hand-in-hand with the play book problem in which peers agree a priori to a series of interactions that will lead to advantageous reputations. For example, a peer may have surrounding peers interact with her over and over again in order to boost her reputation. Then, after being well-reputed on the network, that peer may abuse her powers to perform malicious actions. In certain RMS implementations, such behavior may go unnoticed. This generally occurs in systems with simple summation or average rating computation engines that consider both positive and negative with the same importance ..

Reputation Lag
Reputation lag is defined as the time between an agent’s malicious action and reflection of that behavior in its global reputation. The recommended solution to reputation lag sacrifices the accuracy of reputation computations for speed  and so it depends on the intentions of one’s system.

Value Imbalance Problem
The value imbalance problem is one in which malicious agents utilize the imbalance of transaction sizes to artificially boost their reputation. For example, a malicious user may perform thousands of small file transfers to build its reputation and then transfer a large file with negative intentions and without paying the consequences.