ThinkAir: Dynamic resource allocation and parallel execution in the cloud for mobile code offloading

Abstract

Smartphones have exploded in popularity in recent years, becoming ever more sophisticated and capable. As a result, developers worldwide are building increasingly complex applications that require ever increasing amounts of computational power and energy. In this paper we propose ThinkAir, a framework that makes it simple for developers to migrate their smartphone applications to the cloud. ThinkAir exploits the concept of smartphone virtualization in the cloud and provides method-level computation offloading. Advancing on previous work, it focuses on the elasticity and scalability of the cloud and enhances the power of mobile cloud computing by parallelizing method execution using multiple virtual machine (VM) images. We implement ThinkAir and evaluate it with a range of benchmarks starting from simple micro-benchmarks to more complex applications. First, we show that the execution time and energy consumption decrease two orders of magnitude for a N -queens puzzle application and one order of magnitude for a face detection and a virus scan application. We then show that a parallelizable application can invoke multiple VMs to execute in the cloud in a seamless and on-demand manner such as to achieve greater reduction on execution time and energy consumption. We finally use a memoryhungry image combiner tool to demonstrate that applications can dynamically request VMs with more computational power in order to meet their computational requirements.

Evidence-Aware Mobile Computational Offloading

“…COMET [5] is another framework for code offloading that offloads at thread level, but introduces a DSM mechanism that allows the device to offload unmodified multi-threads. ThinkAir [3] addresses scalability by creating virtual machines (VMs) of a complete smartphone system in the cloud. However, since the development of mobile application uses annotations, the developer must follow a brute-force approach to adapt his/her application to a specific device.…”

“…The conversion tool transforms all methods that fulfill offloading requirements by wrapping each task (or method) 3. http://code.google.com/p/maven-android-plugin/ into a container, which retrieves its execution details from a descriptor during runtime (see Sec.…”

“…The container is named with the name of the task, while the original task is renamed by adding the prefix "local". Similar instrumentation is followed in [3], [6], [8] with the main difference in our approach being the inclusion of a descriptor that governs the execution of tasks based on the community analysis. Besides equipping the app with offloading mechanisms, the conversion tool integrates the device with crowdsensing mechanisms, i.e., mechanisms for collecting context information and sharing it with the wider community.…”

“…Several platforms for providing offloading support have been proposed in the literature with the main difference being the level at which offloading is enabled. Examples of proposed mechanisms include annotation [2], [3], thread-migration [4], thread-migration through data shared memory (DSM) [5], and various others [6], [7], [8].…”

“…The key novelty in EMCO is the use of crowdsensed evidence traces to characterize the influence of different contextual parameters and other factors on offloading decisions [14]. Contrary to existing solutions, which either rely on static code profiling performed on individual devices [2], [3], [4], [5], [6], [11], [15], [16], [17] or on parametrized models that consider a handful of parameters such as network latency, remaining energy, and CPU speed [9], [18], the use of crowdsensing enables EMCO to quantify and characterize the effect of a wide range of parameters and how they vary over execution contexts. EMCO models the context where offloading decisions are made is through simple dimensions that are easy to scale, and determines optimal dimensions using an analytic process that characterizes the performance of offloading based on contexts captured by the community.…”