Collaborative streaming in mobile social networks

Abstract

Mobile social applications have emerged in recent years. They explore social

connections among mobile users in a variety of novel scenarios, including

friend finding, message routing, and content sharing. However, efficiently supporting

resource-demanding delay-sensitive streaming applications on the mobile

platform remains a significant challenge. Research on such topics will naturally

widen the usage of mobile social applications. The solutions to the challenges

will provide suggestion on many related work. It is interesting and valuable to

explore the system performance and users’ experience in such scenarios.

Furthermore, users’ concern about social network is also significant to develop

a mobile social network application. It is important to detect users’ strategies

to communicate with others. That influences the network topologies and

provides biased connections. The strategy consists of various of aspects, most of

which are the user preference and user social attributes.

Focusing on this meaningful research field, we study collaborative VoD-type

streaming of short videos among small groups of mobile users, so as to effectively

exploit their social relationships. Such an application can be practically set in

a number of usage scenarios, including streaming of introductory video clips of

exhibition items to visitors’ mobile devices, such as in a museum. We analyze

users’ behavior strategies based on their social preference and social attributes.

We design SMS, an architecture that engineers such Streaming over Mobile Social

networks. SMS constructs a collaborative streaming overlay by carefully

inspecting social connections among users and infrastructure characteristics of

Bluetooth technologies. To improve the performance, we analyze the scatternet

structure of Bluetooth technology and propose appropriate scatternet structure

in our system. We evaluate our design based on prototype implementation on

the Android platform, as well as on a large emulation testbed. The results obtained

indicate that we are able to achieve a well-performed streaming system in

a mobile social network.