What are the differences and pros and cons between random read write and sequential read write?

Random and sequential reads and writes are the two input and output modes of memory. The stored data occupies space in the disk, for a new disk, the operating system will write the data files to the disk in turn, when some data is deleted, the data will be vacated The original storage space, for a long time, continue to write and delete the data, there will be a lot of scattered storage space, it will cause a large data file to be placed on many discontinuous storage space, when reading and writing some of this part of the data, it is random reading and writing, and the head should constantly adjust the position of the track. Reading and writing data in different locations is much more time-consuming than sequential reading and writing in a contiguous space. When the computer is turned on and when a large program is launched, the computer has to read a large number of small files, and these files are not stored continuously, but also belong to the range of random reading.

Improvement method: Do disk defragmentation, merge fragmented files, but then there will be fragmentation to cause disk read and write performance degradation, and it can't solve the problem of random access to small files, which is just a symptom. A better solution:

Replace the electronic hard disk (SSD), the electronic disk is greatly improved for reading and writing random data because it eliminates the head movement of the mechanical hard disk.

For example, the random read delay of the SSD is only a few tenths of a millisecond, while the random read delay of 7200rpm is about 7 milliseconds, and the 5400rpm hard disk is as high as 9 milliseconds, which is reflected in the performance of the power on and off speed. <>

Random read/write or sequential read/write is not a choice for the user!

In simple terms, when it comes to accessing large files, it's usually access to memory.

A logically contiguous area on a block. Let's say a file is stored on a hard disk and occupies a large contiguous sector.

Then, when reading or writing, it will be written in the order of the sectors.

Conversely, if you want to read and write several files (e.g. a directory), but each file is very small, even a few tens of bytes. Then the storage space allocated for these files must be discontinuous. For example, file 1 is assigned to sectors 1 and 2, while file 2 may be assigned sector 12032 12033.

As a result, new sectors need to be located frequently when reading and writing. in HDD.

This is called "seeking", and it wastes a lot of time. This method is called random read/write. This means that it is not possible to "infer the address" of the object that needs to be accessed in the future from the address of the object that was accessed before. Sequential read/write access addresses are usually sequential.

Random read/write performance reflects the memory addressing capability, while sequential read/write capability reflects the maximum throughput of the memory.

The two are related, but they are not the same thing!

For example, if an M2 hard disk is connected to an inlet, the maximum throughput will be significantly reduced due to the change in interface rate, but it will have little impact on random read and write performance. This is due to the fact that addressing capability is less dependent on interface throughput.

Suppose there are 1 to 1000 transactions.

Situation 1: Now to read the 1000th stroke, the sequential way to read and write is to start reading the first stroke and find the 1000th stroke; Random reads and writes are an operation that quickly finds the 1000th stroke.

Scenario 2: To find the data containing "abc", read and write sequentially or start reading from the first transaction and find the 1000th transaction; Random reads and writes are operations that quickly find the data of "ABC".

Random read/write occupies a lot of space, but the speed is fast and convenient for data processing, while the sequential read and write speed is slow, the space is small, and the data processing is complex.

Suppose there are 1 to 1000 transactions.

Situation 1: Now to read the 1000th stroke, the sequential way to read and write is to start reading the first stroke and find the 1000th stroke; Random reading and writing is through the operation, and the branch is quickly found to the 1000th stroke.

Situation 2: To find the data containing "abc", read and write sequentially or start from the first pen and find the 1000th bit; Random reads and writes are operations that quickly find the data of "ABC".

From the above, it can be seen that random read/write and sequential read/write have different technical methods, resulting in differences in execution efficiency.