嵌入式系统中大吞吐量数据读写模块的设计与实现

发布时间：2018-03-10 00:08

本文选题：数据读写　切入点：嵌入式Linux　出处：《华中科技大学》2012年硕士论文　论文类型：学位论文

【摘要】：数据在磁盘上的读写却时刻面临着各种问题。除了对数据读写的稳定性和准确性的要求外，现在许多场合所面临的大量数据的快速读写调度和读写过程中容易出现的突发性故障的处理等都对当代数据信息的管理提出了全新的挑战。只有将数据读写过程的快速稳定、准确无误、故障及时处理等要求相结合才能使各个产品的品质提高到一个新的层次。为此，针对数据读写的这些较高的需求研究了在嵌入式系统中一个处理大吞吐量数据存取磁盘的问题。为了能够更好的实现将图像压缩数据存储到指定的磁盘上，并且能从该磁盘上读取所需要的数据，设计实现了这套基于嵌入式Linux系统的数据读写调度的应用程序。在经过裁减的Linux系统支持下，分别由数据写入、数据的查找与读出、磁盘管理这些模块来共同实现快速稳定的数据读写。而数据读写的几个难点主要是在于如何高效的将前端FPGA处理打包的数据写入到指定的硬盘保存；此外另一个难题是在出现数据读写错误的情况下如何处理。为此在数据写入模块中设计了共享内存，以提高效率，而在磁盘管理模块中，日志为数据读写出现错误时提供可靠的日志记录信息，以便出现错误后能尽可能的及时有效的恢复数据，同时日志还可以为用户在查询数据时提供相关的检索关键字。在程序完成后经过一段时间验证与测试，实际的运行中可以看到该程序能够较为快速准确的处理大量的前端数据，并能有效的将这些数据写入到硬盘中；同时也能便于用户查找并读取所需要的数据。但是在错误处理上还是存在一定的漏洞，有时出现错误后程序还是无法处理，，整个程序的稳定性还有待提高。
[Abstract]:Reading and writing of data on disk is always faced with various problems. In addition to the requirements for stability and accuracy of data reading and writing, Nowadays, the rapid reading and writing scheduling of a large number of data and the processing of the sudden failure in the process of reading and writing put forward a new challenge to the management of modern data information. Only the rapid stability of the process of data reading and writing is the only way to improve the speed and stability of the data reading and writing process. The quality of each product can be raised to a new level by combining the requirements of accurate and timely fault handling. Aiming at these higher requirements of data reading and writing, the problem of processing large throughput data access disk in embedded system is studied. To better implement storing image compression data on a specified disk, and to read the required data from that disk, The application program of data read-write scheduling based on embedded Linux system is designed and implemented. With the support of reduced Linux system, data is written, searched and read out, respectively. Disk management module can realize fast and stable data reading and writing together. The difficulties of data reading and writing mainly lie in how to write the packaged data of front-end FPGA processing to the specified hard disk efficiently. Another problem is how to handle it in the event of data reading and writing errors. To this end, shared memory is designed in the data write module to improve efficiency, while in the disk management module, Log provides reliable logging information for data reading and writing errors, so as to restore data as soon as possible and can also provide users with relevant search keywords when querying data. After a period of verification and test after the completion of the program, we can see that the program can deal with a large number of front-end data quickly and accurately, and can effectively write these data to the hard disk. At the same time, it is also convenient for users to find and read the data needed. However, there are still some holes in error handling, sometimes the program can not be handled after errors, and the stability of the whole program needs to be improved.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP311.52;TP368.1

【参考文献】