Current solid-state drive technology is doomed, says Microsoft Research

Flash storage has its advantages, such as packing a lot of capacity in a smaller footprint than the traditional hard drive, and the ability to access that data quickly. That’s not going to last, say researchers with the University of California, San Diego and Microsoft Research: 2024 will be the year of reckoning for flash and solid-state drive technology.

Using current chips as a basis, researchers set out to gauge the state of flash technology [PDF] overall. They found that latency and data errors increased as drive size increased. These issues worsened to the point of making the drive too unstable somewhere around 16TB, which the researchers say we will reach sometime in the middle of the next decade.

Making matters worse, the speed advantage that SSDs now enjoy — a common reason to chose the technology over traditional hard drives — is expected to disappear. By 2024, latency will increase by as much as 2.5 times over current rates, the study says.

“While the growing capacity of SSDs and high [data access] rates will make them attractive in many applications, the reduction in performance that is necessary to increase capacity while keeping costs in check may make it difficult for SSDs to scale as a viable technology for some applications,” researchers conclude.

This is definitely a roadblock that looks unavoidable, but there are plenty of technologies in the works that could take the place of flash storage. One possibility is 3D memory, a technology that has been around for the better part of the last decade. 3D seems to be the future in memory, and there are several companies currently working to make it a reality.

3D is not available commercially yet, so that is not an immediate solution to the problem. One thing that is positive here though is the time we have to find a solution. 2024 is a long ways away and 16TB is a lot of space, it will be quite a while before this becomes a problem. Who knows, by 2024 either magnetic RAM or phase-change memory might be mature enough to replace flash.

That said, even the most basic computer users are requiring much more storage space then ever before. You can blame this on digital media and faster broadband, allowing us to download to our hearts content. We need to store that somewhere.

Of course with cloud storage all the rage these days, this storage problem may become irrelevant for the consumer.

Are you using an SSD yet? If not, why not?

Source

Apple might buy Anobit, making a considerable investment in flash

It was revealed today that Apple might purchase Anobit, a fabless semiconductor company. The deal has yet to be confirmed by either party but it would apparently be in the range of $400-5ooM, making it a serious investment for Apple and an aggressive move into the flash storage industry.

Anobit isn’t a company that most people are familiar with, even those who follow technology, but apparently their products are already used in Apple’s iPhone, iPad, and MacBook Air (a banner on Anobit’s sites seems to confirm this). The company doesn’t make flash memory — they leave the fabrication up to others — what they do is improve upon flash technology. Anobit’s MSP (Memory Signal Processing) is designed to lower the cost of flash, improve longevity, and maximize performance. How this happens is pretty technical, but it involves flash-optimizing algorithms that work within the limitations of the flash– be it SLC, MLC, or TLC — to improve performance.

Functionally, this setup works through an MSP-powered flash controller, like the MSP2025. A bit of searching about the MSP2025 only turns up coverage about its announcement back in April but nothing about what products it’s used in. It was noted that the MSP2025 was the industry’s fastest controller designed for use in tablets and mobile phones (666MB/s) and that it could work with 20nm and smaller NAND. The ability to work with sub-20nm NAND as well as to improve performance of more affordable types of flash — namely multi-level cell and triple-level cell — means better flash at lower prices and more capacity for less money.

ExtremeTech analyst Joel Hruska had this to say on the matter: “Triple-level cell flash allows for larger, cheaper SSDs, but manufacturers are concerned about longevity and write cycles. Anobit claims that their MSP (Memory Signal Processing) technology can bring TLC flash up to current consumer specs and solve these problems. If Apple is looking to acquire the company, its probably as a means of increasing available storage at better price points.”

At this point the deal is still a rumor and information about Anobit is largely limited to what they provide on their site, as well as what is secreted away in the bespectacled heads of any number of engineers. What we do know is that Apple is consolidating both its suppliers and companies that they see as able to give them a competitive advantage. Former hardware-focused purchases, like that of PA Semi and Intrisity, haven’t had entirely transparent effects on Apple, but we can be confident that these multi-million dollar purchases are yielding untold benefits, while robbing competitors of access to that technology.

Calcalist, via Techcrunch

Source