OCZ and Indilinx Collaborate On New SSD Garbage Collection Scheme - HotHardware
OCZ and Indilinx Collaborate On New SSD Garbage Collection Scheme

OCZ and Indilinx Collaborate On New SSD Garbage Collection Scheme

A common concern with the current crop of Solid State Drives is the performance penalty associated with block-rewriting. The flash memory used on today's SSDs is comprised of cells that usually contain 4KB pages that are arranged in blocks of 512KB. When a cell is unused, data can be written to it relatively quickly. But if a cell already contains some data--no matter how little, even if it fills only a single page in the block--the entire block must be re-written. That means, whatever data is already present in the block must be read, then it must be combined or replaced, etc. with the new additional data, and the entire block is then re-written. As you can surmise, this process takes much longer than simply writing data straight to an empty block.

This isn't a concern on fresh, new solid state drives, but over time, as files are written, moved, deleted, replaced, etc., many blocks are a left holding what is essentially orphaned or garbage data and their long-term performance degrades because of it.

To mitigate this problem, virtually all SSD manufacturers have incorporated, or soon will incorporate, garbage collection schemes into their drives' firmware that actively seek out and remove the garbage data.

   

 
OCZ Vertex Series SSD

OCZ in combination with Indilinx, for example, is poised to release new firmwares for OCZ's entire line-up of Vertex Series SSDs that performs active garbage collection while the drives are idle to restore performance to like-new condition, even on a severely "dirtied" drive.

We just got our hands on a 120GB Vertex SSD that has been flashed with the new garbage collecting firmware, and wanted to give you all a glimpse as to how it works...
  


New Condition


Dirty Drive


After 5-min Idle


After Long Idle

 
What you see pictured above are four ATTO Disk Benchmark runs performed on the Vertex Series SSD. The first run was performed on the drive in brand new condition. As you can see, performance is high and consistent once the transfer sizes exceeds 64KB. The second run, however, was performed on the drive in a "dirty" state, and as you can see, performance suffers dramatically with transfer sizes below 1MB. This is due to the performance penalty associated with block re-writing.

The third benchmark run was performed after letting the Vertex drive idle for about five minutes though, and performance has nearly been restored to like-new levels, save for a few MB/s drop in the middle transfer sizes. Finally, after letting the drive idle for about an hour, performance has been totally restored.

We haven't done extensive testing just yet, but early indications are that OCZ new garbage collecting firmwares add even more value to what were already popular SSDs. Kudos to OCZ and Indilinx for putting in the effort to enhance the long-term performance prospects of their drives.

0
+ -

this is great "preventative maintenance",

my only concern is about how this effects the life of the cells?

0
+ -

I'm pretty sure this has little effect on the wear leveling algorithms that already are at play for the drives. There's a fairly calculable, finite number of operations that can happen on each cell, so it's easy to know if something would be affected. We'll ask the question though. Good one.

0
+ -

Ditto p3nguin. One of the things I like least about SSDs is their lifespan-- this would seem to work to reduce that, though by how much I don't know.

Of course, motor HDs have a lifespan too, and I've burned out more than one.

0
+ -

I think everyone intiially had/has concerns about SSD lifespans vs. traditional rotating hard drives, but manufacturer's stated MTBFs far exceed hard drives, and those numbers aren't arbitrary.

0
+ -

Well if the SSD's out now can still function and work in 20 years, then i might buy one. As it is the standard mechanical hard drive works just fine and the only time it will fail is when a part inside of it does. I have scsi drives from 15 years ago still working in an old box, and heck i even have 17 year old IDE drives still going, and an even older 5.25 inch HDD still running. Granted mechanical hard drives have hit a wall at transfer speeds atm, and SSD's are a great way to get away from this wall, but for long term reliable storage, mechanical hard drives are the way to go. They are proven, and till SSD's get close to 5 years under their belt of continuous use and not crap out, then you might see SSD's start to replace the good ole hard drive.

0
+ -

Function and work for 20 years? Really? I mean, that's nice that you have IDE drives that are 17 years old, but I can't really see the point of keeping a sub-GB hard drive when I can buy a 2GB flash key at Target for $10 now. Old hardware can be cool and everything, but it really serves zero purpose except as a hobby. 99% of people don't care if an SSD still works in 20 years because it will be utterly eclipsed in capacity and capability by something that is much cheaper.

0
+ -

Oh wow, now that sounds like a LOT of fun!

RT

 

0
+ -

IF these garbage-collection algorithms show their value in the more extensive testing I hope we shall see published here soon and can thus eliminate the performance degradation that hitherto has plagued SSDs and IF no other show-stoppers pop out of the closet, then only one thing can hinder their replacing HDDs - price ! But if and when that comes down....

Henri

0
+ -

I think the garbage collection would only erase cells that would HAVE to be erased in order to use them again anyways. I can see why that would minimize any degrading of cell lifespan.

0
+ -

I think its the same as using thier "wiper" utility, only the drive auto runs it on itself. Running wiper too often will decrease the life of the drive if i remember correctly.

0
+ -

^^^

Actually works completely differently!

How about instead of posting mis-info you know nothing about, you actually go find the correct answer before you mislead a whole bunch of people.

This 'should' be the best GC system we'v seen to date. With this not even TRIM would add any performance benefit.

0
+ -

OCZ once again responds to a need. This will greatly improve overall performance for Vertex Series SSD's for all users whether they migrate to Win 7 or not! My hope is that something similiar can be done for the older Jmicron controller drives so that they can have the same advantage! Maybe a simple firmware update so that the old drives whould tell Win 7 that they are in fact SSD's would do the trick!!!

0
+ -

Garbage collection - great performance - but for which filesystems?

I'm betting that FAT/FAT32/ExFAT/NTFS are all supported, but what about UN*X filesystems?

Or have I missed some really clever way in which the SSD controller identifies a block as "unused" by the filesystem?

Login or Register to Comment
Post a Comment
Username:   Password: