SSD Drives are Defragmented?

Question : Everyone knows that you should defragment the disks periodically to magnetic plates in order to maintain performance at optimum levels.

The same operation is rather detrimental to the unity that SSDs should not be subjected to reorganization as not to compromise the reliability and service life. Many computers, including a Dell XPS27 all-in-one of which are in the possession, use a large magnetic disk flanked by an SSD!

In these cases, the defragmentation should be performed? The defrag utility or the Intel Rapid Storage implement specific functions for these cases in order to rearrange only the magnetic disk drives without involving the SSDs?

Answer : To understand the origin of the problem is necessary to analyze the method used by the software in the Intel Rapid Storage SSD cache management. In configuration’s hybrid hard drives / SSDs, such as the one implemented in the computer of the player, every time the user logs on to a block of the hard disk, it is copied into the cache SSDs. The next time the same block is required, the Intel Rapid Storage retrieves it from the cache SSDs instead request it from the mass memory slower. This mechanism is repeated until the cache fills SSDs. When this condition occurs, each read operation directed to a block that is not yet present in the cache SSDs will result in the elimination of the block that has not been requested for a longer period of time.

This is the most common policy for the management of cache memories and is known as the LRU (least recently used). One detail that you need to keep in mind is that this type of caching operates at the level of blocks and not files. It is possible that, in a file of several GB, the data that is accessed regularly are all contained in a particular section. In this case, the algorithm will only cache blocks most used, leaving the rest of the files only on the magnetic disk.

According to the strategy described above, a possible defragmentation operation can have two undesirable effects: if the software that manages the cache SSDs is not able to distinguish the accesses performed by Defrag would lead each block that is touched by the reorganization to enter the cache, with the inevitable output of other blocks. As a result, the cache will be at the end of the operation to contain unnecessary material and will require a new learning period to restore optimum performance.

But, even in the case where the system of cache management Ssd to restrict itself to ignore the operations of reading / writing performed during defragmentation, this will alter the correspondence between the positions of the blocks and their copies on the hard disk inside the SSD drives, with the need to rebuild the cache anyway.

Given the pattern of operation just described, we can confirm that the hybrid configurations, there is still a need to defragment the hard mechanic to ensure optimum performance when accessing data that can not enter into the cache memory. The reorganization operation inevitably involves a temporary degradation of the efficiency of the cache memory, until they are reported within the unit data SSD longer used.

Since it is not possible to dispense with the rebuild of the cache, instead of letting the unit SSDs is gradually cleared by the algorithm LRU replacement block is preferable to clear its contents and start again from scratch.

This condition can be obtained by simply disabling the caching SSDs before defragmentation and reactivating it soon after. In this way, it will also have the certainty of maintaining full efficiency of magnetic disks without reducing the operational life of the mass storage solid state. The slight decrease in speed of operation, with the inevitable reconstruction of cache SSDs, you will still be largely offset by the improved performance of the magnetic disk.