Storage is no longer the performance bottleneck Storage is no longer the performance bottleneck
New memory tech in Solid-State Drives means primary storage is both faster and cheaper than ever. Storage is no longer the performance bottleneck

A few short years ago, the storage subsystem of any given computer system was the slowest component, unable to parse data as quickly as either RAM or processors. As a result, any waiting the system did was down to those storage components not serving up the required data quickly enough.

Hard drives with spinning platters were the culprits, as they were bound by the laws of physics. The speed at which they could spin, and the rate at which data could be written and retrieved from their platters, had to be capped. Today, most consumer-grade performance hard drives spin at 7200rpm, and in the server world the fastest they ever spun was 15000rpm.

The urban legend around why that is said that if they spun any faster, the edges of the platters would break the sound barrier and shatter. The real reason is far less sexy: pushing platters to go faster would require far too much power to sustain. That would have a knock-on effect when it came to the life span of the drive, and of course power-efficiency ramifications for any systems they run in. Not to mention they’d also take up more physical space.

At the height of hard disk popularity, the only other component that could store data was Random Access Memory (RAM), but it could only do so temporarily – remove power, and the stored data disappeared. While that data was in the RAM, however, it could be parsed at speeds orders of magnitude faster than any hard drive could manage.

It was a natural evolution, then, that hardware developers would look to find a way to make the far-faster RAM store data on a more permanent basis. That research led to what’s known as “non-volatile” memory – memory chips that would retain data even with the power off. This memory has seen several iterations, and today the most commonly-used type is called Flash memory, which is fast and can store a lot of data. You’ve seen it in things like memory cards for cameras, smartphones, and memory sticks, and more recently, also in solid-state hard drives (SSDs).

Flash memory is considered “solid state” storage because of its lack of moving parts, consisting of memory cells packed side by side on a printed circuit board (PCB). While Flash memory proved plenty fast – today’s high-performance SSD-based laptops boast transfer speeds of up to 2200MB/s – it also very quickly ran out of space for more memory cells, and thus its maximum capacity was limited.

When scientists couldn’t get any more memory cells onto the PCB, they solved the problem the same way that city planners did when cities ran out of room: they went up. And that’s how 3D NAND memory was born – instead of packing more memory cells into a 2D space, scientists packed them on top of one another. By going up, they dramatically increased Flash memory capacity and dropped the per-gigabyte cost of SSDs built on the technology while maintaining – and even improving – its performance.

So now, there’s absolutely no excuse for slow systems of any sort. SSDs have replaced hard disks as primary storage, and the most modern ones are fast, affordable and high-capacity thanks to developments like 3D NAND technology.

Tarsus has a longstanding relationship with Intel, and we are well-positioned to assist you with your business hardware requirements. Please contact our Intel Account Managers for a solution tailored to your needs.

No comments so far.

Be first to leave comment below.

Leave a Reply