Network support for energy efficient wireless communications
MetadataShow full item record
In mobile devices, the wireless network interface card (WNIC) consumes a signi cant portion of overall system energy. One way to reduce energy consumed by a device is to transition its WNIC to a lower-power sleep mode when data is not being received or transmitted. This thesis investigates network layer support for for energy e cient wireless communications. We propose a set of client-centered techniques to conserve energy consumption for well-known mobile applications, namely web browsing and TCP downloads. The basic idea behind our network layer energy-saving techniques is that the client actively tracks connections, shapes tra c if necessary, predicts packet arrival time, keeping the WNIC in high-power mode only when necessary. For web browsing with concurrent connections, our technique tracks the connections to identify intervals in TCP streams where putting the WNIC in sleep mode during those intervals is safe and pro table. When upper applications are large le TCP downloads, our technique increases the amount of time that can be spent in sleep mode by shaping the tra c.. Our techniques are compatible with standard TCP and do not rely on any assistance from the server, a proxy, or IEEE 802.11b power-saving mode. We demonstrate the e ectiveness of our techniques by running comprehensive experiments under both emulated environments and the real Internet. We compare our results with regular TCP, PSM, and BSD whenever possible. Results show that our technique combines the performance of regular TCP with nearly all the energy-saving of PSM during web browsing, and we save more energy than PSM during client think times. Over an entire web browsing session (downloads and think times), our scheme saves up to 21% energy compared to PSM and incurs less than a 1% increase in transmission time compared to regular TCP. In the case of large le downloads, results show that compared to baseline TCP, our scheme saves over 50% energy in the best case with a transmission time increase under 8% and on average saves 27% energy with a transmission increase of 20%.