A cheaper alternative to DRAM just took a step closer to enterprise data centers as IBM unveiled a way to make it more dense.
PCM (phase-change memory) is one of a handful of emerging technologies that aim to be faster than flash and less expensive than DRAM. They could give enterprises and consumers faster access to data at lower cost, but there are challenges to overcome before that happens.
Density is one of those, and IBM says it’s achieved a new high in that area with a version of PCM that can fit three bits on each cell. That’s 50 percent more than the company showed off in 2011 with a two-bit form of PCM. Greater density lets IBM squeeze more capacity out of what is still a pricey technology.
PCM has shipped in a few products, but on a relatively small scale. This iteration could do the trick to make PCM a broader success, according to Jim Handy, analyst at research firm Objective Analysis. So far it’s remained more expensive than DRAM, so there hasn’t been much of a reason to use it, Handy said.
PCM works by changing a glass-like substance from an amorphous to a crystalline form using an electrical charge. Like NAND flash, it keeps storing data when a device is turned off, which DRAM can’t do. But PCM responds to requests for data more quickly than flash: In less than one microsecond, compared with 70 microseconds, according to IBM. It also lasts longer than flash, to at least 10 million write cycles versus about 3,000 cycles for an average flash USB stick.
Three-bit PCM could be used as a faster tier of storage within arrays, including all-flash arrays, so the most-used data gets to applications faster. It could also take the place of a lot of the DRAM in systems, cutting the cost of technologies like in-memory databases, Handy said.
Consumers may benefit from the technology, too. For example, a smartphone that has its operating system stored in three-bit PCM could start up in just a few seconds, IBM said.
The sweet spot between memory and flash storage is a popular place to be. Intel and Micron are targeting it with 3D Xpoint, which uses stacked layers of media. Other emerging technologies include RRAM (resistive RAM), MRAM (magnetoresistive RAM) and Memristor.
Like 3D Xpoint, three-bit PCM could break through to large-scale success through the backing of a chip maker, Handy said. Intel is working to make 3D Xpoint work with x86, while IBM is paving the way for PCM with its Power architecture. That kind of buy-in is critical for large-scale production, which is what drives the cost of new technologies down.
IBM added two features to make the new form of PCM possible. One is a way to adjust for so-called “drift,” which can gradually degrade the memory’s ability to store the right values. The other counters the effects of heat on PCM so it can run reliably in normal system temperatures.
Another piece of the puzzle is how to get the data to the processor without slowing down the data on the way. IBM’s betting on CAPI (Coherent Accelerator Processor Interface), a high-speed protocol it uses in Power-based servers. CAPI runs on top of the PCIe physical interface.
IBM isn’t predicting when three-bit PCM will be in mass-market systems, partly because the company doesn’t make memory and will have to find a partner. (It’s cooperated with SK Hynix on PCM in the past.) But it might take two to three years for large-scale availability, the company said.