A method of reading files in disk array at high speed based on fpga

Abstract

The invention provides a method for reading files in a disk array based on FPGA at a high speed, comprising: FPGA sends a file reading request to the disk array; the disk array feeds back a response to the file reading request; FPGA obtains the file size and applies for a corresponding storage space, and constructs an FPGA storage The address notification is sent to the disk array; the disk array sends an FPGA storage address notification response to the FPGA; the disk array performs packet processing on the file to be read to construct a file content frame and sends it to the FPGA; When responding, the file content response frame is fed back; after the disk array receives the file content response frame, it judges whether the sequence number is consistent with the sending sequence number, and if not, resends the corresponding file content frame to the FPGA; the disk array receives the last frame of the file After the content response frame, send the file content completion frame to the FPGA; FPGA sends the file content completion response frame to the disk array. The invention can improve transmission speed and transmission reliability, and reduce FPGA design complexity.

Description

technical field [0001] The invention relates to the field of cloud computing, in particular to a method for reading files in disk arrays at high speed based on FPGA. Background technique [0002] With the development needs of cloud computing and big data processing, field programmable logic array (FPGA) has been more and more used in these fields due to its flexibility, programmability and rich high-speed communication interfaces. At the same time, because the Ethernet switch technology is relatively mature and the cost is relatively low, it has become the first choice for communication between the FPGA and the disk array. However, with the expansion of scale, the high-speed file communication between FPGA and disk array has become one of the bottlenecks in the field of cloud computing. [0003] Under the existing operating system and network environment, most applications require reliable communication, and the TCP / IP protocol has become the main choice for contemporary ne...

AI Technical Summary

Problems solved by technology

However, when applied to a high-speed local area network built with disk arrays and FPGAs, the TCP / IP protocol has three main problems: (1) The throughput is severely reduced in a lossy network. When there is instantaneous network congestion, forwarding errors or optical fiber When there is a problem, TCP / IP cannot distinguish short-term network damage, and the recovery time of TCP / IP is also long; (2) The buffer size of TCP / IP is fixed, and the size of the sender's window is limited by the size of the receiver's buffer. Once a dynamic data packet has not received a response, it will have a strong impact on the transmission of the entire network; (3) The design of the TCP / IP protocol is relatively complicated

In addition to the TCP / IP protocol, the UDP protocol is a commonly used connectionless communication protocol. However, the transmission service it provides is unreliable and the corresponding protocol overhead is relatively large.

Images