在資訊發達科技進步的現代，人們日益仰賴電腦等行動設備以及各種資訊儲存裝置。機密資料及資訊安全如何妥善的加密及防竊已成為重要的議題，重要的個人資料以及機密文件若無妥善加密保管，就容易遭受有心人士竊取並做不當的利用。如何有效的保護重要機密資料，將資料做有效的加密以及需要使用時可以快速的解密資料，是現今資訊安全最重要的一環，如何利用現有的硬體設備讓AES演算法?有效?讓龐大的機密資訊加以妥善加密及保護。本文探討藉由圖形處理器(Graphic Processing Unit, GPU) 在CUDA(Computer Unified Device Architecture, CUDA)架構下執行利用GPU多核心的優勢進行平行運算對區塊加密(Advanced Encryption Standard, AES)演算法做加速，再與以一般傳統的CPU上執行一般AES演算法加解密運算的效能做比較。並且藉由實驗所得的數據，驗證兩種設備的效能高低。而從這些實驗中我們得到了一些研究結果，利用AES區塊加密演算法的過程將分割利用，分別將相同的資料不同資料量在CPU與GPU上面做加解密運算，不同的資料大小在做加密解密的運算CPU和GPU比較出各優缺點，大量資料同時進行加密時GPU平行化的運算優勢讓加解密速度越快。 In the modern times of advanced information and progressive technologies, people are more and more dependent on computer, mobile devices and all kinds of information storing devices day by day. How to properly encrypt confidential information and information security and make them secured against theft has became an important issue. If people don’t properly encrypt and take care of their important personal information and confidential documents, they would easily be stolen and be used at improper place. How to effectively protect important confidential information, effectively encrypt them and quickly decrypt while needed is the most importance in nowadays information security, How to make AES Algorithm more efficient and properly encrypt and protect enormous confidential information by existing hardware devices?This article applied GPU (Graphic Processing Unit) to conduct parallel operation and accelerate AES Algorithm using GPU’s multi-core advantage under the framework of CUDA, and then compared it with the efficiency of conducting encryption and decryption algorithms of general AES Algorithm under traditional GPU. To verify efficiency degree of these two devices by data that received from the experiment. We have received some research result from these experiment; using the process of AES Algorithm and divide it to use, using the same data (but different data volumes) to conduct encryption and decryption algorithms on CPU and GPU respectively; compared different data volumes that that have been conducted encryption and decryption algorithms on CPU and GPU and found their advantage and disadvantage. GPU’s parallelizing operation advantage would make the speed of encryption and decryption faster while conducting a large amount of data at the same time.
