It will be an important problem to monitor the cloud computing for supporting it better, which provides the solution of computing performance adequately. As the increasing of the data, to access the monitoring data accurately and rapidly becomes a limit. It can be improved the efficiency of inquiries for massive data by creating suitable index. In the paper, a storage model was suggested for the large-scale data, and a Two_Layer Index called TB tree (Time B -Tree) which was based on the model is given. The time_piont and time_region access methods were provided on the index, and a good performance was proved by the experiment.