Augmented Smartphone Applications Through Clone Cloud Execution PDF Manual

Smartphones with Internet access, GPS, sensors, and var- ious applications are recently seeing explosive adoption. The Apple iPhone [2], Blackberry smartphones [3], and the Google Android phone [1] are a few prominent ex- amples. In a slightly more advanced capability bracket also lie mobile Internet devices (MIDs) such as the Nokia N810 [7] and Moblin-based devices [6] that provide a richer untethered Internet experience. With popularity, such devices also see new applications by a broader set of developers, beyond the mobile sta- ples of personal information management and music play- back. Now mobile users play games; capture, edit, anno- tate and upload video; handle their finances; manage their personal health and “wellness” (e.g., iPhone Heart Moni- tor [16] and Diamedic [15]).

However, with greater appli- cation power comes greater responsibility for the mobile execution platform: it is now important to track memory leaks and runaway processes sucking up power, to avoid or detect malicious intrusions and private data disclosure, and to manage applications with expensive tastes for high- volume data or advanced computational capabilities such as floating-point or vector operations. Solutions for all these advanced capabilities have been known and are in (fairly) common practice in traditional desktop and server platforms; this is, after all, why smart- phone users expect to apply those solutions to their mobile devices. Alas, such solutions tend to be expensive when cast to mobile architectures. The hardware capabilities of those devices are similar to those of the desktop PCs of the mid-1990’s, many generations of hardware and software behind (see Table 1 and contrast to Table 2).

For example, anti-virus software operates by perform- ing frequent complete scans of all files in a file system, and by imposing on-access scans on the virtual memory contents of a process, including memory-mapped files. On a smartphone, even if the user were patient enough to wait until such a CPU- and I/O-intensive scan were over, she might still hit memory limits or run out of battery power. It only gets worse if one considers tools like taint- checking [23] for data leak prevention, floating-point and vector operations for mathematical or signal-processing applications such as face detection in media, etc. In this paper we (re)discover an opportunity that might overcome these concerns. On one hand, laptop, desktop and server resources are abundant, ubiquitous, and contin- uously reachable, as ensured by cloud computing, multi- core desktop processors and plentiful wireless connec- tivity such as 3G, UltraWideBand, Wi-Fi, and WiMax technologies. The disparity in capability between such computers and the untethered smartphone is high and persistent.

On the other hand, technologies for repli- cating/migrating execution among connected computing substrates, including live virtual machine migration and incremental checkpointing, have matured and are used in production systems [9, 10]. We capitalize on this opportunity here by proposing a simple idea: let the smartphone host its expensive, ex- otic applications. However, do so on an execution engine that augments the smartphone’s capabilities by seamlessly off-loading some tasks to a nearby computer, where they are executed in a cloned whole-system image of the de- vice, reintegrating the results in the smartphone’s execu- tion upon completion. This augmented execution over- comes smartphone hardware limitations and it is pro- vided (semi)-automatically to applications whose devel- opers need few or no modifications to their applications.

Download Augmented Smartphone Applications Through Clone Cloud Execution PDF Manual


Related E-Books:

    None Found

May 9, 2010 | Posted in Electronic Manual, Smart Phone

Comments

Leave a Reply