This paper introduces a distributed secondary control scheme for achieving current sharing and average voltage regulation objectives in a DC microgrid. The proposed scheme employs a dynamic diffusion algorithm (DDA) instead of the consensus algorithm to enable distributed communication among converters. To help understand DDA, the relation of DDA and other diffusion algorithms is discussed in detail and its superiority is shown by comparison with diffusion and consensus algorithms. Furthermore, considering the discrete nature and different sampling time of the digital controller and communication network, a z-domain model of the entire DC microgrid is established. The influence of communication and secondary control parameters on the system stability is investigated. Based on the established model, the tolerable communication rates are obtained. Real-time simulations conducted on the OPAL-RT platform validate the effectiveness of the proposed scheme, showcasing its advantages in terms of convergence speed and stability.