With the recent advances in cloud computing and the capabilities of mobile devices, the state-of-the-art of mobile computing is at an inflection point, where compute-intensive applications can now run on today's mobile devices with limited computational capabilities. This is achieved by using the communications capabilities of mobile devices to establish high-speed connections to vast computational resources located in the cloud. While the execution scheme based on this mobile-cloud collaboration opens the door to many applications that can tolerate response times on the order of seconds and minutes, it proves to be an inadequate platform for running applications demanding real-time response within a fraction of a second. In this chapter, the authors describe the state-of-the-art in mobile-cloud computing as well as the challenges faced by traditional approaches in terms of their latency and energy efficiency. They also introduce the use of cloudlets as an approach for extending the utility of mobile-cloud computing by providing compute and storage resources accessible at the edge of the network, both for end processing of applications as well as for managing the distribution of applications to other distributed compute resources.
TopIntroduction
Recent developments in mobile computing have truly empowered human users, as mobile computing can augment cognitive capabilities dramatically, e.g., through voice recognition, natural language processing, machine learning, augmented reality, and decision-making (Satyanarayanan et al., 2009). With recent advances in mobile devices, coupled with the technological advances in wireless and cloud technologies, computationally intensive applications can now run on devices with limited resources such as tablets, netbooks and smartphones using the cloud remotely as an additional computational resource.
Although different definitions exist in the literature (Dinh, et al, 2011; Fernando et al., 2013), we define mobile-cloud computing as the co-execution of a mobile application within the expanded mobile/cloud computational platforms to optimize an objective function. A typical objective function is the application response time, where the goal is to minimize the objective function. Expanding the application computational resources beyond the mobile is necessary for applications where the objective function cannot be minimized sufficiently by the mobile platform alone (e.g., real-time face recognition), as well as for applications that rely on data not stored on the mobile device. In mobile-cloud computing, it is crucial to provide the user seamless, transparent and cost-effective services as mobile devices rent computing, storage, and network resources from the cloud in order to process and store a vast amount of data (AWS, 2012; Microsoft, 2012; Google, 2012).
We define application cost as an example objective function that quantifies the fees charged by Cloud operators, such as Amazon Web Services, during the execution of the application. For example, Amazon charges for compute-usage per hour per CPU instance, which implies increasing application costs as the required amount of computation increases. Similarly, cloud operators charge for the usage of database instances, such as Microsoft SQL Server. Table 1 shows some example mobile-cloud applications and their computational/storage demands, as well as their application response-time sensitivity. While applications requiring higher computational and storage resources might cost more during operation in a Cloud platform such as AWS, certain response-time sensitive applications, such as the Battlefield application described in Table 1, might tolerate this increased cost due to their need for low response time. Notice that Cloud operators charge less for compute-resources with lower response time guarantees. Specifically, while AWS charges nothing for Micro instances with no response time guarantees, it charges a small amount for the Small instance, and significantly higher for the Large instance, which is a dedicated CPU instance. By the preparation of this document, the AWS pricing for these instances ranged from $0.10 to $0.40 per core per GHz per hour (AWS, 2012), where the unit price decreased with a higher core-count commitment (i.e., number of cores available to an instance). This implies a rich variety of options when executing mobile-cloud applications. The choice of the Cloud CPU instances depends on the application priorities listed in Table 1.