移动端加密算法并行化研究

【作者】 高明；

【导师】 程相国；

【作者基本信息】 青岛大学 ， 计算机科学与技术， 2017， 硕士

【摘要】 近年来,随着智能手机的发展,人们的生活越来越依靠手机,随着国外某些著名品牌手机频频爆出信息漏洞,手机在成为人们连接世界的一扇窗的同时也带来了许多安全隐患。信息安全问题不仅仅是个人隐私问题,更是一个信息化国家的安全保障。目前手机的计算性能已经有了很高提升,多媒体性能成为了移动端的重要参数,但以GPU(图形处理器)为代表的多媒体资源在处理加密、解密等通用计算过程中基本处于半闲置状态。在PC端,GPU为了实现图像渲染实时性,浮点计算能力和数据带宽已经超过了CPU(中央处理器),并且非常适合进行大规模并行计算,国内外学者已经利用GPU进行加密解密算法的研究。相比于PC端,移动端起步较晚,针对于Android平台利用GPU同时保证原有的跨平台的编程技术,从2012年才开始有较为成熟的高性能密集计算框架——RenderScript,目前已经有学者基于此进行移动端通用计算的研究。在Android平台上,本文首次选用RenderScript框架并结合PC端的并行优化方法,对移动端加密算法的并行化进行了研究,主要工作有以下几个方面:研究了DES和AES的实现算法,针对DES的加密过程更多的面向bit操作,对算法中最重要的S盒替换算法进行并行优化,AES以简洁、高效、安全和原则性的设计更适合在移动端进行并行优化,优化了加密算法中的轮变换。针对椭圆曲线密码体制需要用到的二进制扩域上平方运算,发现计算过程中的每个字节之间不存在相互依赖,具有并行计算空间,利用其运算特点,结合串行查表算法,可以在移动端使用并行架构进行编程实现。最后,设计并实现了上述优化方法,并对并行优化的结果进行分析,检验该并行优化的效果。分析结果表明:移动端GPU可以用来加快加密算法的实现效率,通过合理的优化算法以减少异构框架中数据传输的占比,充分利用异构框架的并行特点,实现移动端加密算法的并行优化,为提高移动端加密算法的加密速度提供了一种新的思路。更多还原

【Abstract】 In recent years,with the development of smart phone,people’s lives increasingly rely on it.Facing some well-known foreign brands frequently mobile phone information loopholes,mobile phones,a window to connect the world,bring a lot of security risks.Information security is not only the issue of personal privacy,but also the cornerstone of information security.The performance of the smart phone has been improved.The multimedia performance has become an important parameter.The multimedia resources represented by the GPU(graphics processor)are basically in the semi-idle state in the process of encryption.On the PC side,in order to realize the real-time performance of the Image rendering,the floating point computing power and the data bandwidth of the GPU have surpassed the CPU(central processing unit).Some domestic and foreign scholars have used GPU for encryption and decryption algorithm research.The mobile side is still younger than PC.For the Android platform,available high-performance dense computing framework,using the GPU and ensuring the original cross-platform programming technology,did not appear until 2012.Some scholars are studying general calculation based on the mobile side.On the Android platform,this paper first uses the RenderScript framework and combines the parallel optimization method of the PC side to study the parallelization of the mobile encryption algorithm.The main work has the following aspects:The DES and AES implementations are studied.The encryption process for DES is bitoriented,and the parallel optimization of the most important S-box replacement process is studied.AES—a simple,efficient,safe and principled design—is more suitable for parallel optimization in the mobile side.Round transformation in the encryption algorithm is optimized.For the elliptic curve cryptography system,the square operation on the binary field is studied.Based on the characteristics of its operation,combined with the serial lookup algorithm,it is found that there is no interdependence between each byte in the calculation process,and the parallel computing space can be realized by using the parallel architecture on the mobile side.Finally,the above optimization method is implemented,and the results of parallel optimization are analyzed.The results show that the mobile GPU can be used to speed up the efficiency of encryption algorithm.By optimizing the algorithm to reduce the proportion of data transmission in heterogeneous framework and make full use of the parallel characteristics of heterogeneous framework,it provides a new way to improve the encryption speed on the mobile.更多还原