Research On Key Technologies Of Customed Processor Of Cipher Application

People invented Cipher algorithm and applications to ensure the safety of information. As the base of security technology, Cipher algorithm has been used widly in information field. The Internet developed quickly and more secret information has been stored on the Internet, hence the cipher algorithm become more and more important in applications like Cloud computing and On-line transaction processing. At the same time, due to the development of communication, hole attack and side-channel attack technology, the platform of cipher compute should be faster, safer and more flexible.The traditional hardware and software computing platforms such as CPU, GPU, FPGA and ASIC have a different support on cipher algorithm performance. CPU has a high flexible performance but low performance of computation rate and safety; GPU has a high of compute rate but low safety and of excessive power dissipation; ASIC has high performance in compute rate and security, but the cost is too high and the flexibility is too low; FPGA is of high security and low power consumption, but the computation of rate is normal and flexibility is too low. Each computing platform for cryptographic applications are met only one or two performance requirements.Based on the basis of deep analysis to cryptography algorithm and its application, we designed two customized processors IPCCP and FUMCP for cryptographic applications and providesd the corresponding compiler. At the same time, we designed a visual programming environment oriented to cryptographic application for the customized processor users to facilitate the development of the corresponding cipher application. We also designed a new type of fast frequent subgraph mining algorithm Top-FFSM for excavation frequently used model in cipher algorithm.The IPCCP obtained 57 speedup compared to CPU and 1.8 speedup relative to FPGA. At the same time, FUMCP obtained 1.5 speedup compared to CPU, but the advantage of FUMCP relative to IPCCP is that it has almost exactly the same flexibility as CPU. The frequent subgraph mining algorithm Top-FFSM we designed obtained better effective than C code analysis in both subgraph coverage and execution cycle reduction rate. The visual programming environment we designed has been shown to fully express various cryptographic algorithm application, while the new encryption and password procedures are highly scalable.