Relax--these things take time.
That, in brief, was the message of Jeff Ravencraft, president of the USB Implementors Forum, when I asked him why it was taking the third-generation version of Universal Serial Bus so long to arrive. Intel and others have been touting the sequel to USB 2 since 2007.
"People forget that historically, there is no chipset company in the world that integrated USB from the get-go. It didn't happen with USB 1, It didn't happen with USB 2, It didn't happen with USB 3," Ravencraft said. I takes "a minimum of two years if not more" for a company such as Intel to build USB 3 support into its the chipsets that accompany its processors, he said.
During that time, the first companies build special-purposes chips to handle USB communications, devices get their first support, industry groups hold plug-fests to iron out interoperability problems, he said. Now, with that process well under way, USB 3 is headed for the mainstream.
The cost of USB chips have dropped from about $7 each more than a year ago to about $1 now for devices that plug in with USB 3 and $2 for "host" systems such as computers into which USB devices can be plugged, he said.
"Every penny counts, but at those prices, companies are going, 'Yeah, we're going to roll,'" Ravencraft said.
USB 2 is ubiquitous, spreading far beyond the Windows PCs where it began to Macs, mobile phones, and countless other electronic devices. USB 3, which goes by the name of "SuperSpeed USB" when a device using it passes the USB-IF's testing requirements, is still in the process of coming to market.
USB 3's sales pitch is seemingly a slam dunk: its 5-gigabit-per-second data-transfer rate is 10 times faster than USB 2's, it cuts power consumption by a third, and it can be used to send 80 percent more electrical current so that phones or other devices can charge faster over USB. Intel and others developing the technology handed it off to the USB-IF in November 2008 to manage its real-world use.
The first USB 3 products to market were external hard drives, where the benefits of higher data-transfer rates are clearer, and add-on cards for PCs to let people plug such devices into machines without built-in support. Now a lot more mass storage devices such as thumb drives are arriving. A total of 165 products have passed USB-IF's certification program. Just today, high-end camera maker Phase One announced its new IQ180 80-megapixel medium-format digital sensor module can use USB 3 to transfer its 80-megabyte photos.
But built-in USB hasn't been racing to the market. Intel chipsets supporting USB 3 likely won't arrive until late this year at the earliest, and 2012 is a very real possibility. That would be well over three years from the spec's hand-off to the USB-IF to full integration into the PC hardware realm.
But new interconnects--especially ones with USB's breadth of adoption and cost constraints--are not easy to engineer. When Intel touted USB 3 in 2007, it said it would include fiber-optic links to reach high speeds. In 2008, that optical ambition then moved laterally to a technology Intel called Light Peak--which conveniently used modified USB connectors. Then the initial version of Light Peak switched to copper wiring rather than optical links.
When it does arrive, USB has some room for growth, though.
First off, Ravencraft said, the protocol itself that governs communications was designed to reach a speed five times faster than USB 3: 25 gigabits per second. USB 3 host controllers, which do a lot of the USB talking, also will work at this rate without needing a redesign, he said.
Second, the USB connectors can accommodate fiber-optic lines.
"If we need to go to optical, we've future-proofed the connector to take optical fiber," Ravencraft said. "If that requirement comes about, it'll be a simple transition. I don't think that time is anytime soon."
USB 3 has plenty of potential uses besides external hard drives and cameras. Syncing music and video between a computer and a media player or phone would go faster. Scanners could transfer images to computers more swiftly. Phones and other devices could get universal chargers--indeed Europe is standardizing chargers on a Micro USB interface. External displays could be attached more easily not just to PCs but to phones as well. And note that USB can use the HDCP copy protection content owners demand and that the USB-IF is working on a new USB profile to make video use better.
And should the USB advocates be complacently tempted to presume that past success is a guarantee of future results, there's a cautionary tale with USB: the wireless transition.
Wireless USB has been finished for years, but it remains largely unknown to most people and products, even as the idea of connecting devices wirelessly with Wi-Fi or Bluetooth becomes well-known.
Even with Wireless USB, though, Ravencraft has some optimism--though he cautions that he's "not trying to overhype it."
The version was hampered by different country regulations governing what wireless frequencies could be used and by the difficulties of authorizing a wireless connection between two devices.
But the connection is easier now, and new UltraWideBand (UWB) wireless communication technology using the 6GHz frequency spectrum, introduced with Wireless USB 1.1, gives him some optimism.
Device makers "now have the capability to design a product that has a worldwide footprint," he said. And there's now a good use for Wireless USB arriving that could start encouraging some adoption: "making a simple connection from a notebook to a TV display in the living room without having to run any cables."
It remains to be seen whether it will truly catch on, though. Ravencraft certainly isn't making any promises.
"That space in wireless--there are so many different ways to skin a cat. Wi-Fi, Wi-Fi PAN, Bluetooth, Zigby, Wireless USB, WiMax, wireless HD--they're all in the same space trying to be both the high and low end," Ravencraft said. "I think there's going to be some thinning of the herd."