With the development of network storage technology, object-based storage (OBS), as the next wave of storage technology and devices, provides the advantages of both file storage and block storage. However, the performance of such network storage hinges on knowing the I/O traffic patterns and optimizing the network for such patterns. Recently the notion of self-similarity has been applied to I/O workload, just like to wide-area and local-area network traffic. In this paper, basing on self-similar stochastic processes, we study the long-range dependencies in data traffic for object-based storage system (OBSS) and discuss the relation between I/O workload and network traffic about self-similarity.