INDAPSON: An Incentive Data Plan Sharing System Based on Self-Organizing Network
The contradiction between dynamic user traffic and fixed data plans has drawn increasing attention in the field of mobile applications. We build a data plan sharing system named INDAPSON to consider a scenario where some smartphone users have surplus data traffic and are willing to help others download data. Virtual credits can be gained as reward, which can be used to ask for future help in downloading.
Fig. 1. One example of download group.
As shown in Fig. 1, our local network consists of two different layers: the Bluetooth-Management Network (BMN) and the WiFi-Transmission Network (WTN). We assume that each of the users has access to Internet via cellular data connection. If client A wants to speed up its download process or borrow others’ data traffic, it starts to construct BMN through Bluetooth interface first. Neighboring mobile terminals with the same INDAPSON applications will be invited to join the BMN automatically. To avoid the bottleneck effect of WiFi, only part of terminals in BMN can be invited to join WTN. For instance, in Fig.1 only B and C are asked to join WTN through WiFi interface. Then the download process is started. Client B and C download parts of the data from the Internet using their own cellular interfaces, and retransmit the data to client A through their WiFi interfaces. A merges the data from the other clients and thus consumes less data traffic and achieves much higher download speed. At the end of the download process, all the clients will upload their download records to a management server on the Internet, and thus the contributions of users are recorded. With the scheme of incentive mechanism applied, B and C can get virtual credits from A as reward.
There are three kinds of clients in INDAPSON: primary user, assistant user, and detected user. Primary user (PU) is the client which initially requests data from internet, namely, A in Fig. 1. Assistant user (AU) is the client which helps primary user to download data from server, namely, B and C in Fig. 1. Detected user (DU) is the client in BMN but not in WTN.
Considering the mobility of mobile users, the construction of BMN is operated during the whole download process. Any new neighboring mobile terminals will be invited to be DUs in BMN before the end of download. Therefore the members of BMN change with time. AUs may also join or leave WTN even if the download process has started. If an AU leaves the WTN, a DU in BMN (if it exists) will be invited to be a new AU, and then new download tasks are assigned to this AU.
Power-saving listening method
Since Bluetooth costs less energy than WiFi, we use Bluetooth interface to build the local management network. To further reduce power consumption, we make the clients in Bluetooth listening mode to be intermittent sleeping: with a period of 60 second, the client listens to the wireless channel for 30 seconds, and for the rest 30 seconds it keeps sleeping. For clients in Bluetooth scanning mode, we use heuristic method: the scanning interval is adjusted according to the results of scan. The time interval will be doubled if the results of the previous two scans do not change and the interval has not reached the upper bound we have set. Otherwise it will be halved before it reaches the lower bound. With this method, our system can save more power in relative static environment, and still be able to adapt to dynamic environment.
Fig. 2. Cooperative download group of Android smartphones.
Fig. 3. Average download rate with different number of phones.
Experiments result shows that our system achieves high download rate, which is almost linear to the number of cooperators in a not too large group.