Android Computation offloading

 Abstract : Since it transfers some time-consuming computation tasks to neighbouring servers, computation offloading is a promising method of enhancing performance and lowering battery usage. Various strategies have been put forth to enable computation offloading, though. Given in this paper is an Android application that enables offloading down to the object level. Supporting such a capability is difficult for two reasons: (1) dynamic execution only makes it possible to gather the execution costs of a portion of the methods, and (2) it is challenging to predict the execution costs of the remaining methods. In this paper we are going to outline a framework for offloading complex mobile application duties to the cloud. by heavy tasks, I mean those whose computation takes more time and computing power. Automatic offloading of demanding computation duties to a stand-alone android virtual machine (cloud) is made possible. To determine whether a job needs to be offloaded or not, a variety of factors are taken into account. These factors include how much power the task is using and how much battery life the mobile device on which the application is operating has left. 

 Keywords: Computation Offloading, Code Analysis, Smartphone, offloading, cloud, android, android , cloud , offloader

Across the globe mobile phones are widely used . And with the growing usage several computation intensive applications are blooming in the market . A excellent service that hosts all of your data, running applications, and other programmes is the cloud. One way to conceptualise a cloud is as an ill-defined collection of online-accessible computers and servers. Scalability and portability of cloud computing have made it a remarkable technology. It has altered how we think about communicating and transporting data. Since most smartphones can handle cloud computing environments, cloud services are also heavily ingesting into mobile networks. There is a growth in these apps as mobile phones become smarter, but the device is still constrained because we are unable to enhance the capacity of mobile phones past a certain point. So as discussed in paper we are switching to Mobile Cloud Computing . In the backend we have a complete support system to overcome these difficulties of efficiently running all these applications involving heavy tasks.Google photos and Apple iOS’s Siri are examples of these code offloading techniques as discussed . With the increasing popularity and a large number of developers developing applications for smartphones, the users of these phones have started using them for high end 3D gaming, to handle their finances i.e. internet banking and as their health and wellness managers (e.g. Eat this, not that app. for android). These new applications could be very resource exhaustive and the phones have a limited memory, computational power and battery life. That’s why it makes good sense to offload the heavy applications to the virtual Smartphone running on the cloud thus saving the actual phone’s precious resources. A number of techniques have been proposed to offload the applications of smartphones to the cloud, including complete offloading of the applications as well as partial offloading of the applications. In these techniques used for offloading the application is partitioned at the binary level and thus making this partitioning transparent for the application developer. But this process is compute intensive itself and to make changes at the binary level of an application needs changes in the application loader also which is difficult as well as leads to security vulnerabilities. Furthermore, in the proposed techniques an application called the application partitioner or offloader needs to be installed on the Smartphone which makes the partitions and offloads the appropriate partition of other applications to the cloud. The application offloader makes the offloading decision for all the applications in the phone and thus become an overhead on the phone’s resources. In this paper we propose a framework for offloading an application partially i.e. only the compute intensive, non-interactive part of an application is offloaded. The partitioning is done by the application developer and the offloading decision is taken by the application itself thus eradicating the need of making changes at the binary level and the need of application partitioner or offloader. Many frameworks have been proposed since then . But most of them are not so convenient for developers working out there.In this paper we are proposing a framework for offloading of computation intensive tasks of applications with the use of an already existing framework as mentioned in paper .We are offloading it to a remote server . The framework does not require any changes to be made in the android device side.The static analysis is done to make the decision making more fast and light than the previous techniques. This framework will empower the application to offload its compute intensive part to the cloud via the internet after analyzing the cost of offloading over the cost of running the application on the phone itself. The analysis will be done using parameters like input size and internet connectivity. The remainder of the paper is organized as follows . In section II the existing technologies related to offloading and in section III we summarize the work done related to the idea we are working on .describes the design of the framework which we are proposing and all the architecture of our framework . The conclusion is presented in 

The cloud can be thought of as a terrific service that houses all of your data, operating applications, and other programmes. One could think of a cloud as a nebulous collection of online-accessible computers and servers. Scalability and mobility are two key features that make cloud computing a fantastic technology. It has altered how we think about communication and data transmission. Due to the fact that the majority of smartphones can support cloud computing environments, cloud services are also heavily consuming mobile networks. Mobile phones with enhanced computational power, connectivity, and a wide range of features are known as smartphones. Briefly said, a Smartphone combines the features of a phone, personal digital assistant (PDA), and a laptop.WWith cellphones' rising popularity and the abundance of developers creating applications for them, many users have begun utilising them for high-end 3D gaming, managing their finances via internet banking, and keeping track of their health and wellness (e.g. Eat this, not that app. for android). The phones' limited memory, processing power, and battery life may make these new applications exceedingly resource-intensive. In order to conserve the scarce resources of the actual phone, it makes sense to offload the resource-intensive programmes to the virtual Smartphone running on the cloud. But this process is compute intensive itself and to make changes at the binary level of an application needs changes in the application loader also which is difficult as well as leads to security vulnerabilities. Furthermore, in the proposed techniques an application called the application partitioner or offloader needs to be installed on the Smartphone which makes the partitions and offloads the appropriate partition of other applications to the cloud. The application offloader makes the offloading decision for all the applications in the phone and thus become an overhead on the phone’s resources. In this paper we propose a framework for offloading an application partially i.e. only the compute intensive, non-interactive part of an application is offloaded. The partitioning is done by the application developer and the offloading decision is taken by the application itself thus eradicating the need of making changes at the binary level and the need of application partitioner or offloader.