复杂网络存储系统的层次化元胞自动机模型-计算机系统结构专业论文.docx

II II Abstract* The typical characteristics of network storage system such as the complex change of environment, discrete storage events, the intricate topology of the network architecture, etc, demonstrate the complexity of network storage system. Object-based storage (OBS) system, with some intelligence and adaptability, can be abstracted as an autonomous dynamic system in discrete spatiotemporal. As an effective approach to study dispersed space-time dynamic system, Cellular automata (CA) is capable to analyze the internal rules of data flow of OBS system in complex network. But the regular topology of CA has significant differences from the real network topology. From the view point of whole architecture of storage network system, a hierarchy extended analysis model called TS-A (Transit-Stub-Automata) is proposed. This analysis model extends cellular lattice to the hierarchy domains network topology, and it defines the storage network as the distributed area storage networks (ASN) that connected by the hierarchy domains. So it can reflect more characteristics of the real storage network topology. Hence, this model is a more effective approach to study and analyze the data flow rule of complex network storage system. Based on the research of traditional storage system and OBS, we use this model to simulate and analyze data transmission behaviors of these two systems respectively. Simulation results show that data object flow has global relativity with the phase transition of storage network. When the congestion happens, by using method of giving the forwarding priority to data objects of each ASN according to the congestion degree of the ASN, the congestion will be reduced in certain sense. Hierarchy extended cellular automata can be utilized to simulate the internal complex behaviors and evolvement rules of complex storage system. Furthermore, it can also provide theoretical support for the complex storage system of next generation network. Keywords：Object-based