| 摘要： |
The existing Internet traffic passive measurement solutions are mainly based on a technical route of downloading traffic dataset and analysis tool from the corresponding distribution site, and then carrying out a local place research based on off-line traffic analysis. However, the issues of massive traffic datasets acquisition and analysis as well as measurement achievement reuse and sharing still need a further consideration. The paper applies the ever upsurging cloud computing paradigm to network traffic passive measurement field in order to address the issues. On the basis of inducing the drawbacks of the conventional technical route, we proposed a novel cloud pattern of passive measurement work and designed an architecture of cloud-pattern based network traffic analysis platform. Furthermore, using the authentic traffic collected at a CERNET backbone (10 Gbps), we implemented a prototype system of the architecture, called IP Trace Analysis System, or IPTAS for short. Combined with IPTAS, the paper elaborates the critical implementations of the architecture and verifies its feasibility and flexibility through IPTAS application instances. |