作者: Songwu Lu , Guan-Hua Tu , Yuanjie Li , Chunyi Peng , Zengwen Yuan
DOI:
关键词: Energy consumption 、 Computer science 、 Throughput 、 Computer network 、 Cellular network
摘要: Leveraging multi-carrier access offers a promising approach to boosting quality in mobile networks. However, our experiments show that the potential benefits are hard fulfill due fundamental limitations network-controlled design. To overcome these limitations, we propose iCellular, which allows users define and intelligently select their own cellular network from multiple carriers. iCellular reuses existing device-side mechanisms standard procedure, but leverages end device's intelligence be proactive adaptive selection. It performs monitoring ensure responsive selection minimal service disruption, enhances carrier with online learning runtime decision fault prevention. is deployable on commodity phones without any infrastructure/hardware change. We implement Nexus 6 leverage Google Project-Fi's efforts test among two top US carriers: T-Mobile Sprint. Our confirm helps up 3.74x throughput improvement (7x suspension 1.9x latency reduction) over state-of-art Moreover, locates best-quality most cases, negligible overhead CPU, memory energy consumption.