CN102446261A - Method for increasing writing efficiency of RAID (Redundant Array of Independent Disks) - Google Patents

Abstract

The invention provides a method for increasing writing efficiency of RAID (Redundant Array of Independent Disks), comprising the steps of: firstly opening up a sufficient cache region in system memory for processing user data, grouping the wrote user data according to the RAID disk group in the cache region and sequentially writing the user data in the RAID disk group so at to increase the writing efficiency of the RAID, wherein for fear of the loss of the user data in the cache region, caused by unexpected power down and power off, the data in the cache region are needed to be protected through a BBU ( Battery Backup Unit) or the data temporarily storied in the cache region because of insufficient memory are needed to be manually wrote in the RAID disk group. By utilizing the method provided by the invention, the data reading and writing time caused by write verification is decreased so that the memory property is increased, that is, the writing efficiency of the RAID disk group is increased.

Description

A kind of RAID of raising writes the method for efficient

Technical field

The present invention relates to a kind of data storage technology that relates to, specifically a kind of RAID of raising writes the method for efficient.

Background technology

RAID, raid-array, be a kind of polylith independently hard disk combine by different modes and form a hard disk groups, thereby provide than the higher memory property of single hard disk and the technology of data security.Yet to RAID random writing data the time,, thereby influence the performance that data write, reduce the efficient of writing of RAID owing to write check has produced frequent reading and writing data.Like Fig. 1, the RAID5 disk array group that RAID is made up of 6 hard disks among the figure after in hard disk wherein, writing data, according to the principle of RAID, has the application data of five data in the hard disk for the user this moment, and deposits checking data in another piece dish.Along with the change of application data in the hard disk, need to revise synchronously checking data.Therefore, at first read five user data in the hard disk, then,, draw the checking data of this moment, and the checking data that will newly obtain write the 6th check bit according to the algorithm of RAID5.Pass through said process; RAID according to common writes mechanism; Revise data of RAID disk groups and need carry out five read operations (reading five application datas in the disk) and secondary write operation (write user data and write checking data); By that analogy, when revising five data, need carry out 25 read operations and 10 write operations.It is thus clear that the reading and writing data that caused by write check still has considerable influence for memory property.

Summary of the invention

The purpose of this invention is to provide the method that a kind of RAID of raising writes efficient.

The objective of the invention is to realize by following mode; At first in Installed System Memory, open up enough buffer areas, be used for processes user data specially, in buffer area; According to the user data packets to writing in the RAID disk groups; Order writes the RAID disk groups with user data then, thereby improves the efficient that RAID writes disk, and step is following:

1) one group of batch data is write the RAID disk groups in proper order,, read five user data that write in batches, calculate check bit and write check disk to write check;

2) revise five user data, carry out five read operations and read user data and six write operations, wherein five times for writing user data, once for writing check bit;

3) the data order is write the RAID disk groups, when the not enough position of user data, just need user data be temporarily stored in the buffer memory, write the RAID disk groups together after wait has new data to write again;

4) loss of user data in unexpected power down and the buffer memory that causes of shutdown; Also need protect the data in the buffer area, or will write the RAID disk groups because of the data that not enough position is temporarily stored in the buffer area through manual mode through backup battery BBU.

The invention has the beneficial effects as follows: owing to reducing the reading and writing data number of times that write check causes, thereby improve memory property, promptly improved the efficient of writing of RAID disk groups, thereby, have good value for applications.

Description of drawings

Fig. 1 is the read-write mode synoptic diagram of traditional magnetic disk;

Fig. 2 is the read-write mode synoptic diagram of the inventive method disk;

Fig. 3 is the power down protection synoptic diagram.

Embodiment

Explanation at length below with reference to Figure of description method of the present invention being done.

If want the data order is write the RAID disk groups; At first need in Installed System Memory, open up enough buffer areas; Be used for processes user data specially, in buffer area, according to (quantity of disk-1) user data packets in the RAID disk groups to writing; Order writes the RAID disk groups then, thereby improves the efficient that RAID writes disk.When the not enough position of user data, just need user data be temporarily stored in the buffer memory, after having new data to write again, wait writes the RAID disk groups together.Loss of user data in unexpected power down and the buffer memory that causes of shutdown also need be protected the data in the buffer area through backup battery (BBU).Also availablely will write the RAID disk groups because of the data that not enough position is temporarily stored in the buffer area through manual mode.

Embodiment

As shown in Figure 2, one group of batch data is write the RAID disk groups in proper order, to write check, read five user data that write in batches, calculate check bit and give it and write check disk.This process has been revised five user data once, has carried out reading for five times the operation of user data and has write user data operation for six times, and wherein five times for writing user data, once for writing check bit.Be to revise the five-digit number certificate equally, sequential write is fashionable in batches, write check and the data that cause are write and will be significantly reduced, thus improved memory property, promptly improved the efficient that RAID writes.

Shown in accompanying drawing, be to revise the five-digit number certificate equally, sequential write is fashionable in batches, write check and the data that cause are write and will be significantly reduced, thus improved memory property, promptly improved the efficient that RAID writes.

Owing to reducing the reading and writing data number of times that write check causes, thereby improve memory property, promptly improved the efficient of writing of RAID disk groups.

Except that the described technical characterictic of instructions, be the known technology of those skilled in the art.

Claims (1)

1. one kind is improved the method that RAID writes efficient; It is characterized in that at first in Installed System Memory, opening up enough buffer areas, be used for processes user data specially, in buffer area; According to the user data packets to writing in the RAID disk groups; Order writes the RAID disk groups with user data then, thereby improves the efficient that RAID writes disk, and step is following:

1) one group of batch data is write the RAID disk groups in proper order,, read five user data that write in batches, calculate check bit and write check disk to write check;

2) revise five user data, carry out five read operations and read user data and six write operations, wherein five times for writing user data, once for writing check bit;

3) the data order is write the RAID disk groups, when the not enough position of user data, just need user data be temporarily stored in the buffer memory, write the RAID disk groups together after wait has new data to write again;

4) loss of user data in unexpected power down and the buffer memory that causes of shutdown; Also need protect the data in the buffer area, or will write the RAID disk groups because of the data that not enough position is temporarily stored in the buffer area through manual mode through backup battery BBU.

Images