Never skimp on an SSD! PSA

[Kuno]

6 hours ago, Lady Fitzgerald said:

I'm really tired of seeing this pathetically absurd argument being used against SSDs.

 

All drives, including SSDs and HDDs, are subject to sudden, irrecoverable data with no warning (all drives, given enough time will eventually fail). Depending on being able to recover data from a failed drive of any kind is idiocy. No matter what kind of drive you use or if you use some kind of redundancy (RAID, etc.), the ONLY way to ensure the safety of your data is for it exist in three, separate places. In other words, BACK UP YOU DATA! Then it won't matter if a drive dies. Btw, there are far more ways to lose data than just to drive failure and only backups can protect one from them.

 

Yes, it is true that SSDs, when they fail, usually fail catastrophically (HHDs can do it, too). Despite that, SSDs are still more reliable overall than HDDs (not to mention faster, etc.).

When worst comes to worst, a hard drive still writes sequentially from the outside edge to the center of the platter(s).  Stripping out the servo data and the spare metadata is hard, don't get me wrong, but it is entirely possible to retrieve data from hard drive platters even in the event of almost total translator destruction because higher LBA's can be found closer to the center of the platter(s) in the correct order.

SSD's by contrast require a logical to physical mapping table to match the physical NAND pages to LBA's, and the garbage collection and wear leveling is continuously moving the data on pages around. Trying to reconstruct the data from the flash chips of a modern SSD after the loss of the mapping table is like trying to do a jigsaw puzzle with 125 million piece (for a typical 500ish GB SSD), when all the pieces are the same shape, and the finished picture looks like analogue static. And that's only if the ECC, XOR and compression algorithms can be cracked.

SSD's are generally more reliable, and are almost always faster than hard drives. But there is too many people out there who assume that that means they will never fail. They do fail, and fail hard. Just make sure you keep a backup, and you'll be fine.

[Lady Fitzgerald]

30 minutes ago, Kuno said:

When worst comes to worst, a hard drive still writes sequentially from the outside edge to the center of the platter(s).  Stripping out the servo data and the spare metadata is hard, don't get me wrong, but it is entirely possible to retrieve data from hard drive platters even in the event of almost total translator destruction because higher LBA's can be found closer to the center of the platter(s) in the correct order.

SSD's by contrast require a logical to physical mapping table to match the physical NAND pages to LBA's, and the garbage collection and wear leveling is continuously moving the data on pages around. Trying to reconstruct the data from the flash chips of a modern SSD after the loss of the mapping table is like trying to do a jigsaw puzzle with 125 million piece (for a typical 500ish GB SSD), when all the pieces are the same shape, and the finished picture looks like analogue static. And that's only if the ECC, XOR and compression algorithms can be cracked.

SSD's are generally more reliable, and are almost always faster than hard drives. But there is too many people out there who assume that that means they will never fail. They do fail, and fail hard. Just make sure you keep a backup, and you'll be fine.

The technology to recover HDD data like you said is beyond the scope of the vast majority of users and farming it out to data recovery specialists typically costs thousands of dollars and comes with no guarantee of success. The newer and larger the HDD, the less likely data will be recoverable. All recovery bets are off if there is physical damage to the platters, such as from a head crash.

 

The best policy for both HDDs and SSDs is to back up your data!