Abstract-Smartphone platforms and applications (apps) have gained tremendous popularity recently. Due to the novelty of the smartphone platform and tools, and the low barrier to entry for app distribution, apps are prone to errors, which affects user experience and requires frequent bug fixes. An essential step towards correcting this situation is understanding the nature of the bugs and bug-fixing processes associated with smartphone platforms and apps. However, prior empirical bug studies have focused mostly on desktop and server applications. Therefore, in this paper, we perform an in-depth empirical study on bugs in the Google Android smartphone platform and 24 widely-used open-source Android apps from diverse categories such as communication, tools, and media. Our analysis has three main thrusts. First, we define several metrics to understand the quality of bug reports and analyze the bugfix process, including developer involvement. Second, we show how differences in bug life-cycles can affect the bug-fix process. Third, as Android devices carry significant amounts of securitysensitive information, we perform a study of Android security bugs. We found that, although contributor activity in these projects is generally high, developer involvement decreases in some projects; similarly, while bug-report quality is high, bug triaging is still a problem. Finally, we observe that in Android apps, security bug reports are of higher quality but get fixed slower than non-security bugs. We believe that the findings of our study could potentially benefit both developers and users of Android apps.

Most block-cipher image encryption schemes based on Chaos theory have independent modules for confusion and diffusion processes. None of the current schemes use chaos theory in the diffusion modules -thus not utilizing the capabilities of chaos to the fullest extent. We can do better: we integrate these mechanisms into a single step, thus making the encryption process efficient. This paper presents three novelties: (a) we extend 2D images to 3D by using grayscale image intensities in 8-bit binary form (b) we embed the diffusion mechanism into confusion by applying the 3D Baker map based confusion algorithm. Thus, the diffusion process is accomplished by a permutation of binary bits in the third dimension, eliminating the need for a separate diffusion process and (c) we extend the proposed method to color images by using the 24-bit color information. Color image encryption is usually performed by encrypting each channel independently and then combining these to get the encrypted image. We demonstrate that with this simplistic approach, decrypting even a single channel would reveal reasonable information contained in the image. In our approach, this drawback is eliminated because of the inherent dependence between the data contained in all the channels, thus highlighting the inherent superiority of the proposed algorithm for color image security.

Many vertex-centric graph algorithms can be expressed using asynchronous parallelism by relaxing certain read-afterwrite data dependences and allowing threads to compute vertex values using stale (i.e., not the most recent) values of their neighboring vertices. We observe that on distributed shared memory systems, by converting synchronous algorithms into their asynchronous counterparts, algorithms can be made tolerant to high inter-node communication latency. However, high inter-node communication latency can lead to excessive use of stale values causing an increase in the number of iterations required by the algorithms to converge. Although by using bounded staleness we can restrict the slowdown in the rate of convergence, this also restricts the ability to tolerate communication latency. In this paper we design a relaxed memory consistency model and consistency protocol that simultaneously tolerate communication latency and minimize the use of stale values. This is achieved via a coordinated use of best effort refresh policy and bounded staleness. We demonstrate that for a range of asynchronous graph algorithms and PDE solvers, on an average, our approach outperforms algorithms based upon: prior relaxed memory models that allow stale values by at least 2.27x; and Bulk Synchronous Parallel (BSP) model by 4.2x. We also show that our approach frequently outperforms GraphLab, a popular distributed graph processing framework.

Many vertex-centric graph algorithms can be expressed using asynchronous parallelism by relaxing certain read-after-write data dependences and allowing threads to compute vertex values using stale (i.e., not the most recent) values of their neighboring vertices. We observe that on distributed shared memory systems, by converting synchronous algorithms into their asynchronous counterparts, algorithms can be made tolerant to high inter-node communication latency. However, high inter-node communication latency can lead to excessive use of stale values causing an increase in the number of iterations required by the algorithms to converge. Although by using bounded staleness we can restrict the slowdown in the rate of convergence, this also restricts the ability to tolerate communication latency. In this paper we design a relaxed memory consistency model and consistency protocol that simultaneously tolerate communication latency and minimize the use of stale values. This is achieved via a coordinated use of best effort refresh policy and bounded staleness. We demonstrate that for a range of asynchronous graph algorithms and PDE solvers, on an average, our approach outperforms algorithms based upon: prior relaxed memory models that allow stale values by at least 2.27x; and Bulk Synchronous Parallel (BSP) model by 4.2x. We also show that our approach frequently outperforms GraphLab, a popular distributed graph processing framework.

No abstract