计算机科学与探索 ›› 2018, Vol. 12 ›› Issue (9): 1407-1419.DOI: 10.3778/j.issn.1673-9418.1710028

• 系统软件与软件工程 • 上一篇    下一篇

MultiOffloading:支持多种资源借调的安卓应用平台

张晨曦1,2,武翔宇3,许  畅1,2+   

  1. 1. 南京大学 计算机软件新技术国家重点实验室,南京 210023
    2. 南京大学 计算机科学与技术系,南京 210023
    3. 微软(中国)有限公司苏州分公司,江苏 苏州 215000
  • 出版日期:2018-09-01 发布日期:2018-09-10

MultiOffloading: Android Application Platform Supporting Offloading Multiple Resources

ZHANG Chenxi1,2, WU Xiangyu3, XU Chang1,2+   

  1. 1. State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing 210023, China
    2. Department of Computer Science and Technology, Nanjing University, Nanjing 210023, China
    3. Microsoft China Co., Ltd. Suzhou Branch, Suzhou, Jiangsu 215000, China
  • Online:2018-09-01 Published:2018-09-10

摘要: 安卓应用数量飞速增长,应用场景不断丰富,但部分安卓应用的运行受到智能手机本身CPU性能不足或传感器功能缺失的限制;同时在使用场景上也面临如何远程控制的问题。将应用内部的部分计算任务转移到远程服务器中计算是一种普遍的解决手机性能不足的方法,但是该方法仍存在服务器连接复杂等方面的问题。提出了MultiOffloading,一个同时支持安卓应用计算任务卸载与转移执行、传感器功能“借用”和触摸控制转移的平台,搭载这个平台的移动设备可以互相发现并建立P2P网络,平台上的移动应用能够“借用”移动设备计算资源、传感器资源和屏幕资源;重新实现和完善已有的计算任务卸载、传感器借用的插桩技术并提出触控转移的插桩实现方案。最后实现了一个示例以说明平台设计的可行性,并选取两个安卓应用插桩展示资源借用的效果。

关键词: 安卓应用, 计算卸载, 传感器借用, 触控转移

Abstract: The number of Android applications is increasing rapidly. However, some Android applications running on the smartphones are limited by the conditions of the smartphones themselves, such as insufficiency of CPU resources or malfunctioned sensors. They also face the problem of remote control due to their richer application scenarios.  Computation offloading is proposed to solve this problem by transferring the complex computation tasks contained in the app to a server to execute and then returning the result back to the app. But there still remain problems such as complicated client-server connecting. This paper presents MultiOffloading, an Android application offloading platform which supports computation offloading, sensing offloading and touch offloading at the same time. The smartphones which have already installed MultiOffloading can find and connect each other to form a P2P network. The applications on the platform can borrow computation, sensor and screen resources from other devices. This paper extends existing computation offloading and sensing offloading approaches and enhances them with the support for touch offloading via program instrumentation. At last it gives a demo to confirm the feasibility of the offloading platform design and chooses two Apps to display the offloading effect.

Key words: Android application, computation offloading, sensing offloading, touch offloading