期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2017
卷号：17
期号：8
页码：28-34
出版社：International Journal of Computer Science and Network Security
摘要：Due to the advancements in computation technolo- gies, apparent in small, faster, and parallel processing capabilities both in hardware and software, there has been a paradigm shift to ubiquitous computing. Most of the computing platforms exist in the form of mobile computing, where portability and round- the-clock accessibility is the main advantage. Mobile Cloud Computing (MCC) helps bridging Cloud Computing seamlessly into a mobile environment to elastically enhance resource utilization in the on-demand requirements for the mobile users and mobile applications providers. The main goal is the provision of a better experience for mobile users by alleviating problems concerning storage, computational power, and battery. This removal of resource limitation occurs in the form of Task offloading. Cloud has the capability to compute faster using the numerous resources available, more storage performance and a constant supply of power compared to mobile devices, hence ��heavy-lifting�� the cloud is the optimal approach to mitigate resource limitations. Many models have been proposed which take certain parameters into consideration for offloading such as the speed of the cloud versus the speed of mobile devices, and the bandwidth available for transfer between the cloud and the mobile device etc. More- over, before offloading decisions need to be made on what are ��offloadable �� elements of the code. This annotation is done either manually by the developers of the applications or automatically through appropriate partitioning tools. CloneCloud, which is the focus of this paper, helps in automatically marking the potential ��offloadable �� blocks in bytecode in a static fashion and then at runtime determines the optimal offloading. CloneCloud uses virtual machine migration (VMM) to transfer mobile application blocks to a cloud server using either 3G or Wifi. This paper elaborates on the framework details of the CloneCloud, the shortcomings found in this model and our proposed solution.
