Article Preview
TopIntroduction
Smartphones are a favorite means of communication among people through out the world. According to the International Telecommunication Union (ITU, 2013), as in early 2013, there were almost as many mobile phones as people on earth (ITU However, smartphones are rapidly replacing ordinary mobile phones. With the booming growth in smartphones, the number of smartphone applications (apps) is also growing rapidly. Among all types of smartphone applications, emails, multimedia streaming, and mobile web browsing are more popular. According to Cisco, mobile web browsing data accounted for 25% and mobile multimedia traffic accounted for 51% of all mobile traffic by the end of 2012. It is predicted that two-thirds of the global mobile data traffic will be multimedia streaming by 2017 (Cisco Systems, 2013).
Basically, web browsing and multimedia streaming applications were designed to be used in desktop computing environment where computers are connected to the Internet via a wired medium. As of today, both web browsing and media streaming applications together account for 75% of the mobile data traffic. When those applications are adapted to mobile computing environment, energy and bandwidth limitations of the mobile devices must be taken into account.
Web contents delivered to end users are undergoing radical changes in the past few years, because most of the apps are available at no charge. Webpages used to be simple, static pages comprising only text and some images. However, modern webpages are dynamic and media-rich. The publishers of those pages rely heavily on advertisements (ads) as a source of income. The ads, in turn, 25 are media-rich and complex for resource limited mobile devices. Completely eliminating ads from webpages is not viable because ads generate revenue and apps are downloaded by users free of charge. When a webpage is displayed on a user’s screen, downloaded are: (i) the core information that a user is actively interested in, namely, news, emails, and video; and (ii) extra information in the form of advertisements. Figure. 1 shows a real example of the type of contents that are downloaded and displayed on a smartphone.
Figure 1.
Sample website showing the core information and some ads (highlighted by ellipses)
In a smartphone environment, where energy and bandwidth resources are limited, displaying ads on the screen is an expensive task. To be displayed, ads consume energy in three stages: fetch, render, and display. Each of those actions requires specific resources: (i) network access, namely, radio interface and bandwidth; and (ii) computations on application processors and graphics processors.
In addition to the extra costs of handling ads, multimedia streaming applications consume much energy. Therefore, it is important to make streaming applications energy-efficient. Similar to the web based ads, streaming applications consume energy in three distinct stages: content delivery, content decoding, and content playing (Naik, 2010). It is important for streaming applications to have an energy-efficient downloader, because the radio modules of a smartphone consume much energy. Since all smartphones use a WiFi interface supporting the IEEE 802.11 power saving mode (PSM) by default, it is straightforward that the PSM will be utilized to minimize the energy consumption during streaming sessions. However, the standard PSM has limited effectiveness for streaming applications, because streaming applications often involve bulk data transmission in a continuous fashion, which would not provide many opportunities for the wireless network interface controller (WNIC) to switch into PSM. Consequently, we show that application designers need to develop their own energy-efficient downloaders. With energy and bandwidth in the backdrop, this paper makes the following contributions: