SSD middleware breakthrough gives up to 300% speed boost

Scorpus

Posts: 2,162   +239
Staff member

Solid state drives are already one of the fastest storage solutions going around, but a new breakthrough in SSD middleware has the potential to make the drives even faster.

Due to the inherent way NAND works, it's impossible to directly overwrite data stored in areas of memory, so when you write to the drive, it writes the data to a different area first before invalidating the old area. This causes drive fragmentation, which is currently solved through a process called "garbage collection" that rearranges data through a series of writes, slowing down the overall writing speed.

A team from Japan's Chuo University have developed a new technology that partially solves this problem; a problem which has been present in SSD technology since the beginning. The solution is complex, relying on a "logical block address scrambler" to reduce the effects of fragmentation while simultaneously reducing the amount of copies needed during garbage collection.

The new technology has a number of benefits that were observed during the team's testing. First and foremost, they managed to increase the write speed of an SSD by up to 300%, but they also saw a 60% reduction in power consumption, as well as a 55% decrease in write/erase cycles that should help prolong the lifespan of drives.

It's also possible to apply the new type of middleware to existing solid state drives, as it doesn't require any modifications to the NAND hardware. It seems unlikely that SSD manufacturers would push out such significant software updates to their old products, but we could see the tech integrated into new drives over the next few generations.

Permalink to story.

 
"It seems unlikely that SSD manufacturers would push out such significant software updates to their old products, but we could see the tech integrated into new drives over the next few generations."

Why not! Why can't they release a driver updated for PC users.
 
"It seems unlikely that SSD manufacturers would push out such significant software updates to their old products, but we could see the tech integrated into new drives over the next few generations."

Why not! Why can't they release a driver updated for PC users.

Because most of the existing SSD products are already operating at maximum throughput supported by SATA-III, so it will make little difference, but surely many support problems.
 
"It seems unlikely that SSD manufacturers would push out such significant software updates to their old products, but we could see the tech integrated into new drives over the next few generations."

Why not! Why can't they release a driver updated for PC users.

Probably because that would mean users wouldn't upgrade aka spend cash on new products which isn't what a business wants :/
 
"It seems unlikely that SSD manufacturers would push out such significant software updates to their old products, but we could see the tech integrated into new drives over the next few generations."

Why not! Why can't they release a driver updated for PC users.

Probably because that would mean users wouldn't upgrade aka spend cash on new products which isn't what a business wants :/

If I buy a new SSD, it will be more for capacity than speed. I store my OS on my SSD, and there's barely enough space as is. That being said, if I can prolong the life and increase the speed of the drive while reducing power consumption, all with a firmware update, I'd do it.
 
Was thinking about getting a larger SSD to replace the 60GB one I have for my OS. Guess I will wait a little longer until one with this new "technology" comes out. I don't expect a 300% increase but any increase would be beneficial.
 
If this is just a better algorithm and easy to update as this article makes it seem like then all manufacturers will want to retrofit their older drives. It makes the company look fantastic and their products fantastic at very little cost.
 
If this is just a better algorithm and easy to update as this article makes it seem like then all manufacturers will want to retrofit their older drives. It makes the company look fantastic and their products fantastic at very little cost.

No if the company is anything like a normal business they will profit from it not give it out or free! Hence why electric cars are not free!!!!
 
"It seems unlikely that SSD manufacturers would push out such significant software updates to their old products, but we could see the tech integrated into new drives over the next few generations."

Why not! Why can't they release a driver updated for PC users.

Because most of the existing SSD products are already operating at maximum throughput supported by SATA-III, so it will make little difference, but surely many support problems.
Not really, it is because they want money! It is a companies dream to have breakthroughs like this one so that they can say, "look, 300% more performance! **big numbers** ***ooooohhh ahhhhhh***" and the consumer says, "wow, what great tech, lets spend another $200 even though I just bought one a few months back!"
 
Was thinking about getting a larger SSD to replace the 60GB one I have for my OS. Guess I will wait a little longer until one with this new "technology" comes out. I don't expect a 300% increase but any increase would be beneficial.

You might as well go ahead and purchase a new one could be years before anything like this is coded for consumers....That's if ya need it...
 
"It seems unlikely that SSD manufacturers would push out such significant software updates to their old products, but we could see the tech integrated into new drives over the next few generations."

Why not! Why can't they release a driver updated for PC users.

Because most of the existing SSD products are already operating at maximum throughput supported by SATA-III, so it will make little difference, but surely many support problems.
Yes mainly support. Bearing in mind, there is still plenty of headroom for 4k random writes so 300% gain there would be magnificent.
 
"It seems unlikely that SSD manufacturers would push out such significant software updates to their old products, but we could see the tech integrated into new drives over the next few generations."

Why not! Why can't they release a driver updated for PC users.

Because most of the existing SSD products are already operating at maximum throughput supported by SATA-III, so it will make little difference, but surely many support problems.

Not true as SATA-III can support up to 6gb/sec throughput, where most SSD drives top out at around 500ish mb/sec
So they actually could improve to be at least 10x as fast before reaching the SATA-III threshold
 
Not true as SATA-III can support up to 6gb/sec throughput, where most SSD drives top out at around 500ish mb/sec
So they actually could improve to be at least 10x as fast before reaching the SATA-III threshold
In sequential reads in particular, SATA-III has been saturated since the Vertex 2. ie. the first sandforce SATA-III drive.

500MB/s is the close to the practical limit. SATA-III is 6gbps (gigabit/s) NOT 6GB/s (gigabyte/s).

The reason the spec is defined in bits per second is that is what goes over the physical channel. There aren't just data bits though - there is overhead on top. The user cares obviously about bytes though - that's their data. Rough rule of thumb for many comms channels is 10bits per byte but in SATA it appears worse (more overhead).
 
"It seems unlikely that SSD manufacturers would push out such significant software updates to their old products, but we could see the tech integrated into new drives over the next few generations."

Why not! Why can't they release a driver updated for PC users.

Because most of the existing SSD products are already operating at maximum throughput supported by SATA-III, so it will make little difference, but surely many support problems.
even if performance wasn't increased, having less wear on the drive would be nice.
 
In sequential reads in particular, SATA-III has been saturated since the Vertex 2. ie. the first sandforce SATA-III drive.

500MB/s is the close to the practical limit. SATA-III is 6gbps (gigabit/s) NOT 6GB/s (gigabyte/s).

The reason the spec is defined in bits per second is that is what goes over the physical channel. There aren't just data bits though - there is overhead on top. The user cares obviously about bytes though - that's their data. Rough rule of thumb for many comms channels is 10bits per byte but in SATA it appears worse (more overhead).

Ah, I have thought 6gbps to be in bytes, not bits. Yeah I see what you mean now, definitely close to the SATA III limits already where it would be 750mb/sec hard drive throughput, minus the overhead
 
Back