Application Research On Network Attached Optical Jukebox

The explosive increase of digital information demands a reliable storage and effective utilization for mass data. As a partial online storage facility in the hierarchical mass storage, Network Attached Optical Jukebox has a very important application value in scientific research, production and many other fields. The key of improving the application of super-big capacity storage instruments is to enhance its application performance.In this thesis, based on the advanced research on the basic model of Network Attached Optical Jukebox, the factors that influence the application performance of Network Attached Optical Jukebox were analyzed, and the approaches and methods for improving the transfer performance were probed. All of them provided a solid academic and experimental foundation for the application of Network Attached Optical Jukebox. The main contents of the thesis include:Firstly, the characteristics of the arrival of user's accessing requests of the Network Attached Optical Jukebox were investigated. Based on it, a queuing model with different priorities was constructed, and the key factors including the priorities of each class were analyzed and calculated. Based on the research on the logical relationship between the user's accessing behaviors, an"Active Restricted Multi-Accessing Model"with time-effectiveness and repetitiveness was built, so that a new model foundation for the improvement of the Network Attached Optical Jukebox application was provided.Secondly, the difference between the Caches in the Network Attached Optical Jukebox and in the magnetic disk was analyzed in the thesis. Based on the"Active Restricted Multi-Accessing Model", the Cache substitution algorithm under the different accessing load conditions was derived for the Network Attached Optical Jukebox. The research work showed that, the operation times to the Network Attached Optical Jukebox can be reduced by the effective load processing mechanism, so that the access capability can be improved, and the performance of the mass storage system was enhanced. Thirdly, aiming at the real application environment of the Network Attached Optical Jukebox, its application capacity was evaluated by the linear neural network algorithm. Based on the research on the Peer-to-Peer System, combining the application characteristics of the Network Attached Optical Jukebox, the Peer-to-Peer System model of the Network Attached Optical Jukebox was created. It showed that, by the Peer-to-Peer model, the Network Attached Optical Jukebox could not only reduce the system responding time, but also enhance the system throughout. Thereby the system accessing load quantity was reduced.Finally, based on the above research work, the software and hardware module structure was designed and implemented by the Network Attached Storage, and the application performance of the Network Attached Optical Jukebox was evaluated via experiments. The testing results showed that, the Network Attached Optical Jukebox with the new processing mechanism had a high application performance in practical use. The mechanism which effectively dealt with the user's accessing behaviors provided a new plan for the networking application of the super-big capacity storage instruments, and made the Network Attached Optical Jukebox with the characteristics of high capacity, high usability, extendibility, and easy to manage.