Today, one of the most important challenges encountered by researchers in Internet of Things subject is how to organize a large number of devices in this environment. The organization should be in a manner that the objects can effectively perform the service discovery process. Recent efforts led to emerge of a new paradigm known as Social Internet of thing or briefly SIoT model. According to this model, objects will be able to reach their goals efficiently through the concept including creating friendships and navigating in a social network of friends. On the other hand, the service discovery time will be prolonged according to the increase the number of objects, the distribution of objects throughout the network, the local and limited knowledge of the objects about their neighbors and the inefficiency to the response to a special request by the objects. Therefore, we propose service based grouping. The goal of adopting this approach is to increase the efficiency of service discovery process within the paradigm of SIoT. Due to the open environment of SIoT, each object can freely get-in and get-out in the network. Therefore, in such a dynamic and distributed environment, each object can share any service or information to the other ones. Consequently, the interaction with these unknown objects without any restriction and control mechanism confronts us with security challenges including free-riding problem and selfish behavioral. Key Words: Social Internet of Things, Trust, Free-Riding, Multi-layer Networks, Centrality Measurement (HITS)
