Content Centric MSNs
The current network architecture in MSNs is location-based, where the content is delivered to users based on their IP. Due to the ever expanding scale of MSNs, there are multiple replicas of the same content which is stored at different locations. For example, two users may keep the replica of the same content downloaded from different sites. As these two replicas own different IPs, they may be recognized as different content, resulting in a mistaken content exchange. In fact, users care more about what this content is than where this content is from. Content centric networks have been advocated to manage contents based on their naming information [4, 5]. The content is requested by sending interests in it from users.
In content centric MSNs, the content is recognized by a hierarchical name containing its content ID. When multiple replicas of the same content are delivered to different users in different communities, these replicas can be identified from the same content. The mistakes of content exchange and the efficiency to manage different replicas can be expected to reduce.
In addition, the content that is frequently requested by users in certain community can be stored at the nearby content cache node, which can reduce the delay for users to obtain the content in this community. Thus, how to design a content-centric MSNs becomes an open problem.