In this work, our research focuses on a design for request distribution and associated security attacks in dense vehicular ad hoc networks (VANET) and also sparse VANET that creates a delay tolerant network (DTN). Generally vehicles are stratified into clusters; we presented a reliability based clustering which has been designed for VANET. Cluster creation is according to complicated clustering metric that considers density of relation graph, link value and also traffic conditions. Since the ones in specific time and location are always affecting with the similar pattern of the direction and also velocity. A vehicle communicates with other vehicles or it's nearest Road Side Unit (RSU), which provides an access for a local cloud for sending appeals. We define the formal security model k-anonymization of privacy preserving aggregated transmission evidence generation (ATEG) in our proposed trust based VANET network (TBVN). It is required that both the individual vehicle velocity and the average velocity of vehicle clusters should be well protected from the semitrusted vehicular cloud and the malicious running vehicles. Therefore, except for the traditional security requirements such as data secrecy and authentication, unique safety and privacy concerns are emergently should be rectified.