The increasing number of vehicles has been causing many problems during the past years such as traffic congestion, environmental pollution and traffic accidents, etc. Recently, the impact of Connected and Autonomous Vehicle (CAV) on traffic system has received extensive interests, which is regarded as a promising driving pattern to significantly improve traffic safety and efficiency. Due to the developing technologies, more information flow topologies may appear in the future, which brings more challenges in the study on CAV. Thus, it is necessary to classify the topologies systematically to address the problem. Currently, some studies have provided certain insights on the influence of information flow topologies from several aspects including the internal stability, scalability, asymptotic stability and robustness etc. However, to the best of our knowledge, the influence of communication topologies on string stability is still unsatisfied and unclear. Thus, motivated by the various communication information topologies and the gap of string stability, this paper focuses on the impact of CAVs on heterogeneous traffic with different communicated information topologies, which are divided into three categories: forward communication topology, backward communication topology and bi-directional communication topology. By using the transfer functions, the string stability conditions and corresponding stable areas are derived under different communication environments. The theoretical results are validated by the numerical simulations, which shows that the backward communication topology has the weakest string stability while the bi-directional communication topology has the strongest string stability for the vehicle platoon. These findings are meaningful for the design and control of CAV system in the future, which will further improve the traffic planning and management.