March 22, 1999
Will USB wither on the vine?
By Bob O'Donnell
After years of waiting, watching, and hoping, the Universal Serial Bus (USB) has
finally begun to make its mark on the personal computer marketplace. USB brings many
benefits to PC users, including hot-plugging, reasonably fast performance, and most
importantly, the capability to upgrade a system without having to worry about interrupt
requests and other PC resource issues. And now that both Windows 98 and MacOS 8.5, among
other operating systems, offer full-fledged support for the burgeoning standard, USB is
even starting to bring some consistency across platforms, which is good for both vendors
and users.
Of course, USB isn't perfect. As I discussed in a previous column, The
Not-so-universal serial bus, there are still occasional compatibility issues with
different types of USB controllers (OHCI, or Open Host Controller Interface, vs. UHCI, or
Universal Host Controller Interface) and USB devices. In addition, power-related issues
are still a big bugaboo, partially because of confusing terminology and partially because
of misunderstandings of what USB can and cannot do -- I discussed some of these issues in
another column The realities
of USB aren't what you were led to believe.
Power-related issues with USB hubs in particular has been a big problem and even led to
a refinement of the original USB standard. The changes in USB 1.1, which was formalized
last fall and is backwards compatible with USB 1.0, are particularly focused on
hub-related issues. (In fact, if you're looking for a USB hub, make sure it supports the
USB 1.1 standard.)
This isn't the only change in the USB spec, however. At last month's Intel Developer's
Forum, the company announced the creation of a standards body and preliminary
specifications for USB 2.0, which is expected to appear in systems by the middle to end of
next year. According to information on Intel's
site, USB 2.0's primary feature will be an increased transfer rate. In fact, the rate
is supposed to increase by a factor of 10 or 20 times, which would give USB a transfer
rate of 15 to 30 Mbs (or 120 to 240 Mbs, as they like to say). More than just improving
performance for existing peripherals, these transfer rates would open up the possibility
of adding a wide variety of high-performance peripherals, including storage devices,
high-end digital video cameras, and more to the bus.
While these rates don't match those for IEEE 1394 -- which currently offers 50 MBs (or
400 Mbs) and is expected to double that in later iterations -- they are close enough that
it's bound to raise comparisons with and questions about the relationship between the two
standards. Until this announcement of USB 2.0, USB was positioned as the low-speed
connection of the future and IEEE 1394 as the high speed option. But now, that distinction
isn't quite as clear.
While there are certainly good technical reasons for improving a standard such as USB,
I have to wonder if these recent USB 2.0 announcements are related to the fact that Apple
(who invented the IEEE 1394, or "Firewire," technology) announced a while back
that they intended to collect royalties from any company using IEEE 1394.
Given Intel's delays in implementing IEEE 1394 support in its chipsets (a development
that was originally promised for the end of 1998) and now the announcement of USB 2.0, it
certainly appears that Intel is backing away somewhat from its IEEE 1394 support. Worse,
the move seems to be more political than technical, because of the fact that Intel
developed USB (and hence doesn't need to pay royalties), but didn't develop IEEE 1394 (the
old NIH, Not Invented Here, syndrome).
I'm certainly pleased to see a useful standard like USB get improved in a way that will
extend its life and bring additional capabilities to end users, but I hope it doesn't come
at the expense of other useful standards, such as IEEE 1394. When it comes to making
computers easier to use, the industry can still use all the help it can get.
©
Copyright 1999, by InfoWorld Publishing Corp., a
subsidiary of IDG Communications, Inc. Reprinted from InfoWorld,
155 Bovet Road, San Mateo, CA 94402. Further reproduction is prohibited.
