Why SSDs seem to present the regressive development?

In 1956, American IBM Company had developed the world’s first computer hard disk which was as big as two refrigerators and has the capacity of 5M. While an ordinary M2 SSD, designed in 2019, is 2.3 millimeters thick and has a capacity of 500 G. Technological upgrading from HDD to SSD is a leap-type development. But why is the iteration of the SSD itself a retrograde development?

A common HDD is made up of a disk rotating 7200 times per minute and magnetic head writing data every few milliseconds. Most of the computer noise you usually hear is caused by the mechanical movement of the hard drive.

SSD
A piece of SSD is made up of controller, cache and flash memory. It takes less than 0.1 ms for the controller to access multiple pieces of data in different locations on the flash memory. Thus the reading and writing speed of SSD hundreds of times faster than that of HDD. The key part of SSD is the flash memory that stores the data, which determines SSD performance. The flash memory was originally generated from a wafer. According to different needs, Chip manufacturers cut out qualified wafer chip and package them, then flash memory called wafer is produced. This is the origin of the flash memory called a wafer. For decades, major chip manufacturers have successively launched four types of chips with different structures.

First generation of SLC

In the SLC flash memory, the smallest memory cell of flash memory can store one bit of data, which is a 0 or 1, so a cell only has two states.

Second generation of MLC
In the MLC, producers have put two bits of data into one cell that a combination of 0 and 1. So a cell has four states.

 Latest generation of QLC
From what we have discussed above, in the QLC flash memory of the latest generation with a memory cell presenting sixteen states. And, the storage capacity is eight times larger than the SLC flash memory. But, all the gifts of fate have already marked the price. Because doubling storage capacity can also lead to reduced flash memory speed and lifespan. Different from HDD, the number of erasing and writing of the SSD flash memory is limited, which is the lifetimes of flash memory. When the number of erasing and writing of flash memory has reached the limit, the flash memory will be damaged and the data store will be lost.

The number of erasing and writing of the SLC flash memory can reach one hundred thousand times. In the QLC flash memory of the latest generation, the number of erasing and writing has decreased to 1000 times. Under the circumstance of same capacity and usage, the theoretical lifespan of the two flash memories is 100 times different. Therefore, with the development of technology, the performance of each generation of SSDs has regressed, which can reduce the cost of SSD so that ordinary people can afford them. Even if the performance is compromised, the lifetime of SSD is still enough for ordinary people to use. In addition, the operating speed of SSD is still faster than HDD. Generally speaking, the problem of SSD is not the short lifetimes but the operating speed will slow down when the SSD is full of data. The problem is related to the process of erasing and writing of flash memory.

Full storage space

We use these lattices to present the storage space of flash memory. Among them, the blue parts are the position in which has stores data. If you delete the profile, the hard disk may not dismantle the data of the space but it will mark it. When you want to store a piece of profile, the hard disk may not write the new data in the marked space but store the data in the blank area, using more storage space. With the increase of reading and writing, the blank zone of hard disk will be used. At this time, if you want to input the new data, you need to clear out these marked spaces. In other words, you need to dismantle the data. But, the useful and useless data will be deleted at the same time when you clear out the data of flash memory. So you need to move the useful data to other spaces and clear out this space and then input new data.

Input the useful content into the other spaces
This process is called write amplification which means more complex steps, much time and more the times of erasing and writing. Therefore, if the SSD is full of many data, the operating speed and lifetimes will be reduced. In order to solve this problem, the manufacturer will choose the other way of regress rather than progress to shorten the capacity of hard disk.

Reserve some blank space
Reserving some blank space can stop users to fill in the hard drives. This is why the SSD capacity we see on the market is often marked with 240G or 480 G instead of 256G and 512G.

Related Articles:

The solid-state disk structure of SSD：controller algorithm ...