Big data computing and analysis can uncover hidden patterns,correlations and other insights by examining large amounts of data.Comparing with the traditional processor,the new types of processors,just like digital signal processor(DSP),FieldProgrammable Gate Ar-ray(FPGA),graphics processing unit(GPU),could improve the speed of data analysis significantly.Heterogeneous multicores systems have be-come the primary architecture as devices are tasked to do more com-plicated functions faster.While,in most cases,these heterogeneous re-sources cannot be utilized sufficiently because the system software is provided by vendors,loaded pre-sale and doesn鈥檛change.The cloud computing offers the capability of distributing infrastructures according to the requirements.We build a cloud-like heterogeneous computing plat-form which including PowerPC,DSP,GPU and FPGA.A task-driven dynamic loading scheme is proposed by making use of the virtualiza-tion and middleware technologies.The system can manage the entire lives of allocating,loading,using,and recovering.Taking this as a guide,a private cloud principle verification system including web application layer,main control layer,and computing service layer is designed and verified,which proves the feasibility of the computing platform.Accord-ing to the test results of web system,the platform can well meet the design intention of acquiring the computing resources according to the task requirements.