The Design And Implementation Of Clustering Module In Scalable Network Attached Storage

With the development of information technology, the enterprises depend on electronic data more than ever before. The enterprises have to solve a series of problems such as data storage, management and usage. In storage field, network attached storage (for short, NAS) products have a lot of advantages, so they are used a lot. They are products designed for a special purpose with low-price and features support for multiple data transmission protocol. Unfortunately, traditional NAS products are based on single computing node. They are hard to extend and not reliable, so the need of round-the-clock service can not be satisfied.This paper will introduce a scalable NAS product. It is a cluster NAS based on General Parallel File System of IBM. With the features of distributed file system and management system, this product provides extention and management operation while working. Scalable NAS management system is based on B/S architecture. The enterprises can adjust the number of computing nodes according to the size of data. This policy can effectively reduce the purchase and operating coses. Based on the scalable NAS product as background, this paper will introduce the ayalysis, design and implementation of the cluster module in this project.Firstly, this paper introduces the background of saclable NAS product and the situation of research in NAS area. The role and functions of the cluster module in this project will be introduced. The product has4-layer architecture. From bottom to up, the order is network layer, data layer, service layer and presentation layer.Secondly, this paper introduces the technology and tools used in the project.Thirdly, this paper introduces the requirement analysis in the project. After that, the design and implementation of cluster module is introduced. Cluster module has4major submodules, the network submodule, the node management module, the privilege management submodule and the performance monitoring submodule. In this paper, function design for each of the submodule, key process of the product is discussed in detail. Finally, this paper makes a conclusion, and points out what the cluster module can be done in the future, such as improving the auto load-balance.