Interview Western Digital (WD) listens more than it speaks. The company's roadmap is rarely revealed and it sails serenely on saying nothing, while flash solid state drives (SSD) challenge hard disk drives and cloud backup squares up to WD's external backup drive business. We don't know whether it's in a state of denial or knows things we don't know and these upstart technologies have less substance than we suppose. Should we get real, or should WD?
WD would have us, politely, get real.
Flash is not the holy grail. It's just another storage medium and it's not going to take over the world. Hard drives store the bulk of the world's online digital data today and they will do the exact same thing tomorrow and the day after that. I paraphrase - this isn't Western Digital (WD) speaking directly. It's my spin on the WD position after talking to Richard Rutledge, WD's SVP for marketing. See what you think.
Richard said: "We see NAND flash having two areas of value. There is very cheap storage, below hard disk drive (HDD) capacity, such as USB flash drives, digital camera cards and mobile phones. You can get an SD card for $19.99 whereas a hard drive costs $50 or more. This flash is a very low budget purchase."
Then: "There is very, very, high-performance flash, the STEC products. In between the value of a hard drive is pretty difficult to beat." He characterises DRAM as the fastest storage with read/write speeds in the nanosecond area (10 to -9 seconds). Flash is in the microsecond area for reading (10 to the -6) but millisecond area for writes (10 to the -3), and hard drives are in the millisecond area for both read and writes. That's three neatly defined tiers.
WD's view is that the SSDs delivered in 2007 were re-purposed camera cards. The technology to speed flash writes - to make multiple flash chips operate concurrently - came from high-end camera cards. Early digital camera users wanted to take pictures quickly one after another, as they could with film, but couldn't with single-chip camera cards. So high-end cards were devised with up to four NAND chips, written concurrently. This led to SSD technology. STEC SSDs now have eight or more chips, hence the Mach 8 name.
How does WD know all this? Because it was one of the original investors in SanDisk in the late eighties and brought controller technology to SanDisk. WD owns some of the original SSD patents and Irwin Federman, SanDisk's vice chairman, is an ex-WD director.
Rutledge also wants us to know that the bulk of shipped netbooks, Asus Eee-type machines, use hard drives and not flash. It's as if he wants us to understand that flash is not walking over any HDD market.
Small form-factor drive developments
What kind of things might WD do to strengthen its HDD offering? Is a 15K rpm Velociraptor feasible? (WD currently ships a 10K rpm 2.5-inch drive called Velociraptor.) "15K is feasible," Rutledge said. "The market for 15K drives is predominantly servers ... We don't do products for every market. SSDs in the enterprise space challenge 15K rpm drives." The impression I took from this is that WD is unlikely to introduce a speeded-up Velociraptor drive.
Could WD supply a SAS Velociraptor? "Today we supply Velociraptor as a 2.5-inch SATA drive. We only announce products when they are available. ... The 6Gbit/s SAS standard is in good shape. The industry will adopt it over the next twelve months." Straining hard to read between the lines here I think we might reasonably expect a SAS Velociraptor.
The focus for WD is not to go head-to-head with flash. NAND products will win in the very low-budget storage market and look strong in the very high-performace enterprise market, where WD does not have products. Flash is challenging in the netbook and notebook space, but HDDs have the better cost/capacity profile here and in desktops and bulk enterprise storage. WD will keep on increasing the areal density of its products so as to retain this advantage.
Rutledge said: "Our industry tends to double capacity on a two-year cycle." There are three underlying technologies: perpendicular writing and tunnelling reading in the head, and the media itself. He reckons WD will be doing more work on the media than the head.
The impression received here is that WD sees no unbeatable technology obstacles to moving on from today's best areal densities of around 400Gbits/sq in to the 1Tbit/sq in and beyond level.
WD is the market leader in externally-attached hard drives for consumers, it being a one-and-a-half billion-dollar revenue earner for the company. Here services like EMC's Mozy and Carbonite are offering automated backup to the cloud. Would WD consider adding a cloud backup service to its external drive offerings?
Rutledge said that once consumers get used to doing backup one way they do it that exact same way pretty much forever. They write to CD or DVD, they backup to an external drive: "There is no one solution. We need to deliver the best backup solution we can. Whoever provides a better experience and peace of mind to consumers will have a good business. People want easy to use backup - KISS - keep it simple, stupid." More than once Rutledge mentioned iTunes as a marker for a good consumer experience.
He said that USB is simple and well-understood, and that: "The Holy Grail backup quest is less to argue about where the data goes and more about how to get more people doing it."
My impression? Don't expect a WD cloud backup service anytime soon.
So ... is WD in denial about flash and cloud storage, or are the flash and cloud proponents stuck on Fantasy Island? Certainly WD does not believe in having feet in two camps. Unlike Seagate it has no SSD product strategy and no i365 cloud services on offer. WD is doing what has served it so well so far, and sticking to its HDD knitting. Knit one, purl one, purl another, and another, and another. Why stop doing what works? ®