54 per Va Si V E Blockincomputing

T he recent appearance of open operating systems and virtual machines optimized for mobile devices has led to new and innovative applications.1 However, although these mobile devices pave the way for a new generation of applications that exploit their intrinsic mobility, proximity to the user, and dependence on handheld devices, they create unique challenges for developers.2 For instance, new mobile devices come with powerful multitasking processors, but their performance and processing capacity differ significantly from those of traditional desktop computers. In this restricted and unknown development environment, developers face new constraints related to not only hardware issues such as processors and memory but also the lack of tools for debugging the different aspects that affect mobile application performance. Mobile device technologies such as Bluetooth, wireless local area networks (WLANs), General Packet Radio Service (GPRS) or Enhanced Data Rate for Global Evolution (EDGE), Universal Mobile Telecommunications System (UMTS)/ High-Speed Packet Data Access (HSDPA), and WiMAX also introduce variables during development that developers working on fixed networks are unaccustomed to. Each technology has different behaviors and is suitable for different contexts. In addition, the simultaneous use of different radio-access technologies is a new challenge for mobile developers. Several existing tools can analyze a mobile application’s resource consumption performance. However, these tools don’t address communication performance from a developer’s standpoint (see the “Related Work in Application Profiling” sidebar) or answer these questions: