Research On Privacy Protection Strategy Against Collusion Attacks In Friend Search Engine

With the development of social networks,people's awareness of the personal privacy data has gradually increased,and the protection of privacy data has become one of the research hotspots.Among them,the friendships between users are the basis of social relationships.The research on its security issues is of great significance for both social network users and service providers.However,due to insufficient personal privacy and security awareness or malicious attackers in the network trying to gain benefits by stealing personal privacy data,users' privacy data will inevitably be leaked.At present,friend search engine,a kind of social application,appears in social networks,which can provide users with the service of querying the list of other persons' friends.A malicious attack against the users' friendships has subsequently appeared.The attacker can obtain the friendships that the target user is allowed to display outside the range by querying through the friend search engine.Existing protection strategies can resist malicious attacks initiated by independent attackers.However,the collusion attack jointly initiated by multiple attackers.Compared with the independent attack,its attack behavior is difficult to be detected,which leads to serious leakage of the privacy of the users' friendships.Aiming at the existing collusion attacks on the privacy of friendships,the paper proposes an anti-collusion attack strategy that combines trust metric and threshold function.The purpose is to protect the privacy of the users' friendships from being damaged by the attackers while ensuring the users' normal query.The paper first proposes a measurement method of trust between users based on users' behavior,analyzes and explains the calculation methods of direct trust,recommended trust,and comprehensive trust.Regarding the trust degree as the characteristic attribute of the users,and using it as the key indicator of whether the user can make a query,the trust threshold is set to classify the requester who inquires the users' friendships.Secondly,by using Shamir secret sharing system,the friendship between two users is transformed into a secret,and the(,9))threshold function is applied.After the secret distributor distributes the secret,the secret reconstruction process is improved to ensure that the participants who meet the query requirements can successfully reconstruct the secret,while the participants who do not meet the query conditions cannot successfully reconstruct the secret even if they obtain the secret fragments.Finally,through experiments,the rationality of the trust measurement method proposed in this paper is verified,and the security verification and analysis are performed against collusion attacks.Experimental results show that the strategy can limit the malicious queries of collusion attackers while ensuring the normal query of users,reduce the probability of successful collusion attacks,and protect the privacy of the users' friendships.