Does the fatal gene of ‘write endurance’ built into flash SSDs prevent their deployment in intensive server acceleration applications? That was certainly a show stopper in 2005, but now, with the capacity and the ‘write endurance’ increase, SSD has made through high-end enterprise storage vendors like EMC.
What the ‘write endurance’ all about at first place? You must understand that the number of write cycles to any block of flash is limited, it was about 10.000 in the early 90’s and today, in late 2009 some vendors claim 5 millions write cycles.
What happens once you have reach the limit? The disk can become unreliable, or at least the block which has been written 5 millions times is unreliable anymore.
What do the SSD vendors do to overcome the ‘write endurance’ issue? The common method is scoring how many times a block of memory has been written to, and then reallocating physical blocks to logical blocks dynamically and transparently to spread the load across the whole disk.
Here is a formula to estimate the SSD endurance limited life:
(write endurance) x (capacity) / (write speed) = the endurance limited life in seconds
You divide by 3600 to get the endurance limited life in hours
You divide by 86400 to get the endurance life in days
You divide by 31536000 to get the endurance life in years
This is purely theoretical as the formula doesn’t take account the way your OS or application behaves. In another word ‘write endurance’ is not anymore a factor you should worry about!