NAND Flash, Intel's Braidwood, and PCs

A few weeks ago, I attended the Flash Memory Summit and posted several blog entries, which generated a few comments. Recently, analyst Jim Handy released a report on Intel’s Braidwood, which is Intel’s 3^rd generation Flash cache for PC architectures. Flash cache is a different way to boost hard-disk drive (HDD) performance. Instead of putting together a bunch of Flash chips to create a solid-state drive (SSD) with HDD-like capacity, a Flash cache holds contents drawn from the HDD that are needed immediately, leaving the rest of the data stored on the HDD alone.

Here are some of the bullets from the executive summary in Handy’s Braidwood report:

• Since a NAND layer will offer near-SSD performance for less than 10% of the price of an SSD, fewer PC buyers will purchase SSDs than do today, even though today’s SSD penetration is only about 1%.
• SSDs will not be the only casualty of this technology. Over time OEMs and end users will find that they get a bigger performance boost for their dollar by adding NAND than by adding DRAM, and the DRAM market’s megabyte growth will decline further, causing revenues to shrink over the long term.
• Once a NAND memory layer is proven to be beneficial it will find rapid acceptance. Objective Analysis expects to see Braidwood adoption sweep the market in only four years.
• No matter how high the acceptance of this technology, its contribution to NAND revenues will remain low peaking at about 7% of total market revenues.
• Software support is key to the acceptance of a NAND layer in the computing hierarchy. Prior generations of Flash cache, including the Hybrid HDD and Intel’s Turbo Memory or Robson failed due to faulty software support, which Intel is addressing with Braidwood.

You can purchase Handy’s full report through his company Objective Analysis here.