Research On Key Technique Of Cloud Cryptographic Resource Pool Scheduling

Cloud cryptographic resource pool which can provide tenants with large-scale cloud cryptooperation services,is an important infrastructure to solve the problem of cloud computing security.Aiming at the problem on resources scalability and service reliability of the existing cryptographic resource pool schemes,a dynamic reconfigurable Field Programmable Gate Array(FPGA)cryptographic resource pool scheme is proposed.Secondly,key technique of virtual FPGA scheduling like workflow prediction,virtual FPGA placement and virtual FPGA migration are studied.Finally,based on the fundamental code,cloudSim toolkit is extended.FPGA entity and virtual FPGA entity are developed for simulation experiment and performance evaluation.Simulation results show that the proposed strategies can both improved the resource utilization and reduce the number of active FPGAs,while guaranteeing the quality of service.The main research contents and innovations of this dissertation are as follows:1.Aiming at the problem on resources scalability and service reliability of cryptographic resource pool,a dynamic reconfigurable FPGA cryptographic resource pool framework is presented.FPGAs are used to provide tenants with crypto-operation services to improve resources flexibility and service reliability.Based on the fundamental code,the cloudSim toolkit is extended and FPGA entity and virtual FPGA entity are developed to simulate and evaluate the scheduling strategies of cloud cryptographic resource pool.2.An extended Auto Regressive Integrates Moving Average model(ARIMA model)with multi-task and multi-stage prediction is proposed to deal with the perceived lag of requirement changes.Extended the model predictions dimensions of ARIMA model,to provide multi-stage prediction on computing capacity requirements of different crypto-operation.It enables cloud cryptographic resource pool system to timely perceive the dynamic change of tenants' cryptooperation requirements.3.Aiming at the problem that static virtual FPGA placement strategies often lead to large number of FPGAs idle,an optimized ant colony optimization(ACO)algorithm is proposed.By giving ants the ability of perceiving resource status and designing a free space,it can not only implement a load balancing among FPGAs,but also improve the utilization ratio of FPGAs in the cloud cryptographic resource pool.4.Aiming at the problem that aggressive migration of virtual FPGAs usually causes an increasing of system communication overhead,a hierarchica virtual FPGAs migration algorithm is proposed.In the algorithm,a global scheduler and a local scheduler are respectively set on the FPGA resource pool and FPGA,to process virtual FPGA scheduling operations on a piece of FPGA as much as possible,which can effectively reduce the times of virtual FPGA migration,and reduce communication overhead of cloud cryptographic resource pool system.