Last Friday, Intel introduced one of the smallest flash-memory-based hard drives on the market. The chip, also known as a solid-state hard drive, competes with similar chips from Samsung, which store data in gadgets such as Apple’s iPod nano and iPhone. But the Intel chip comes with a standard electronics controller built in, which makes it easy and inexpensive to combine multiple chips into a single, higher-capacity hard drive.
The move highlights Intel’s effort to establish itself as a leader in flash-memory chips and to make them a replacement for the bulky and conventional magnetic hard drives that store data on most of the world’s computers. Smart phones and so-called ultramobile computers will require some kind of dense, durable storage system in order to bring the power of desktop computers to handheld devices.
Since it found its way out of the lab in the late 1990s, flash memory has revolutionized consumer electronics. Because flash-memory chips are smaller, more rugged, and more energy efficient than magnetic hard disks, they have been the ideal replacement for hard drives in handheld devices such as MP3 players, and even in some high-end laptops. Flash is a solid-state memory technology, which means that it has no moving parts and stores data using silicon transistors like those found in microprocessor chips. Because it uses microprocessor technology, it also roughly follows Moore’s Law, the prediction that the number of transistors on a chip doubles about every two years. For processors, this means that they get faster, but for flash-memory chips, it means that data storage doubles. And the market has responded to flash’s burgeoning capacity: in 1999, the flash-memory market was nonexistent, but in 2007, it amounts to $15.2 billion.
At a press event, Don Larson, the marketing manager of Nand products at Intel, showed off the new chip. Called the Z-P140, it’s about the size of a thumbnail and weighs less than a drop of water. It currently comes in two- and four-gigabyte versions, which are available to manufacturers for use in handheld devices. The first products featuring the new chips will be available in January.
Since the new solid-state drive has standard control electronics built in, it can be combined with up to three other Intel chips that don’t have controllers, for a maximum of 16 gigabytes of storage, says Troy Winslow, flash marketing manager at Intel. While that may not seem like a lot compared with the 160-gigabyte hard drives in desktop computers, Larson pointed out that two gigabytes is enough to run some operating systems, such as Linux, along with software applications. The chip’s electronics also allow it to work well with Intel processors, which make it useful for the ultramobile-PC market. And by 2010, Larson said, Intel expects to be able to cram 64 gigabytes of storage into a piece of silicon about the size of the new chip.
Flash has drawn criticism because its memory cells, which hold the electrical charges that represent data, tend to wear out quickly. But Winslow says that in the new chips, a memory cell can have data written to it and erased from it up to 100,000 times. And to ensure that no single cell gets overused, the chips have “wear-level” algorithms programmed into them, which evenly distribute reading and writing. Thus, flash memory could start to show signs of wear in about five years, depending on how it is used. In addition, the static electric field that holds charge tends to degrade over time; data losses in this case could occur after about 10 years.
Researchers at Intel and other companies are looking for the next solid-state technology that could replace flash. Winslow says that Intel is currently testing phase-change memory, a type of memory in which the crystal structure of a material changes in response to heat; particular orientations of the crystal correspond to 1s and 0s. Phase-change memory has many of the benefits of flash, such as its ruggedness and small size. But data can be written to it many thousands of times faster than it can to flash.