Implementation Of A File System Based On FPGA And Its Application

The rapid expansion of the various data in present-day society makes people enter the era of big data. Big data is the data to which a single addressable unit can hold uncertain or arbitrary size data. It is very large scale and complex data sets, traditional data storage management mechanism will face many new challenges when dealing with the data that frequently reach PB, EB or the ZB-level. It is the major topic of this study to seek an effective way to meet the requirements of highly concurrent read and write of big data, demand for high efficiency storage and access of the mass data and scalability and high availability of storage capacity through the research of bottom layer storage of big data.Corresponding to the characteristics of big data, implementation of file system on FPGA is proposed in this study by analyzing from such levels as bottom layer medium storage of data, file system management and interface access mode and so on. The implementation process is introduced through concrete examples and the problems encountered are analyzed and solved, verifying the feasibility of the scheme, which is specifically as follows:First, design and optimization of the bottom-layer ATA host controller. Re-design and optimization is carried out based on the existing ATA controller in our laboratory, improving the read and write functions of Ultra DMA, summarizing the read and write mechanism of bottom-layer hard disk in DOS, Windows and Linux operating systems and summing up the main parameters of underlying hard disk interface and then re-packaging interface which makes it easy to connect to the upper-layer file system.Second, FPGA Implementation of Ext 2 hardware file system.(1) Through related source code and based on an in-depth understanding of the storage principles of FAT32 and Ext2 file systems, the commonalities of file systems are extracted and the basic operation process of files are summarized: create files, read files, write files, delete files, update documents, etc.;(2) Hardware file system and the bottom-layer ATA host controller are connected directly; the various functional modules of Ext 2 file system are described in VHDL and the code is debugged and simulated;(3) Two specific implementations of file system are put forward:state machine and sequencer engine, which are elaborated on according to relevant characteristics.Third, implementation of multi-mode and remote-access client.(1) Remote-access is carried out through WEB browser under B/S mode: design front-end HTML page, compose CGI programs, add Boa server to the PetaLinux operating system in FPGA, compile the file system library, etc.;(2) The remote-access process of files is achieved by composing Socket applications under C/S mode.In the specific scheme of achieving file system with FPGA in this study, the approach of implementing the basic data structure of file system with RAM is put forward; the character-search function is applied to hardware circuit, realizing file name matching; hard disk command control module reasonably controls the commands sent to the hard disk, greatly reducing data read and write time.From the various angles of data bottom-layer storage, the scheme of achieving the file system on FPGA proposed in this study completes read and write operations on data and documents without CPU scheduling, realizes the direct interaction between the user data cache and the hard drive, saving resources and shortening the time of file processing. The scalability of the system can be greatly increased by separating the file system and the hard disk controller as an independent resource node, which can provide references to meet the needs of big data and achieve efficient allocation of resources and services.