【作者】 刘东；

【导师】 陈华钧；

【作者基本信息】 浙江大学 ， 计算机应用技术， 2012， 硕士

【摘要】 遥感技术是当今人们获取空间信息的一种重要手段,而由于技术的发展和需求的提高,遥感系统要处理的数据量越来越大,对大数据传输的要求也越来越高。网格是一种资源共享与问题解决的有效手段,它可以实现了计算、存储等资源的跨域协同,从而可以对大规模的数据处理与传输提供很好地支持,因此网格技术是解决遥感系统中所面临的大数据传输的理想方法。为了满足遥感系统中对大数据传输方面的需求,我们针对遥感影像数据基于网格架构提出了一个传输系统：DartFTP。本文介绍了该系统的总体设计与具体实现,并对关键技术做出了详细地讨论。遥感影像数据体积较大,并且格式特殊,而遥感影像数据处理系统由许多彼此独立的子系统构成,这些子系统的工作方式和数据访问方式都各不相同,该传输系统基于主要针对这些特性进行设计,充分考虑到与原生产系统的结合方式与交互接口。在此基础上,本系统主要有如下特色功能：多协议支持,对原有系统中存在不同种类的文件服务器提供支持；副本传输,可以同时从系统中多个节点上的副本开始进行传输;服务质量控制,提供传输系统中各个层次的对服务质量的支持。更多还原

【Abstract】 Remote sensing is one of the most important methods to obtain spatial information now. While the technology is advancing quickly, and the needs from different departments also increasing rapidly, the remote sensing system may have to deal with larger scale of data in shorter time. Grid computing technology is an effective means for resource sharing and problem solving. It can achieve cross-domain collaboration with computing resources and storage resources. The grid computing architecture can support large-scale data processing and transfer, and it is ideal for solving the problem the remote sensing system has incurred.In order to support large-scale transfer problem in the remote sensing system, we designed a transfer system called DartFTP based on the grid architecture. This article describes both the design and the implementation of this system, and we would talk about the key points in detail. Remote sensing data is often in large scale and with a special format, also the whole remote sensing system is constituted by several independent subsystems, which each often has its own workflow and specific way for data access. Our transfer system is well designed for these characteristics, and it also fully takes the way of interaction with original system into account. Based on this, our transfer system has the following features:multiple transmission protocol support, which provides a way to access different file servers in existing system; transferring with replicas, which means obtaining data from different replicas on different servers at the same time; QoS support, which means the system could provide a way to control the quality of service on different layers on the system.更多还原