Letters to the Editor.
It seems that CTR generally does a good job of providing timely, yet accurate information. So I was quite surprised when I noticed a number of errors or at least misimplications in an article that appeared in the October 1999 issue ("USB 2.0: A Storage Bus Whose Time Has Come", page 32).
My job isn't based on either of these technologies and I wish them both well. However, I am quite bothered by one-sided articles that do not fairly present the facts.
The author suggests that IEEE P1394 (also known as FireWire, iLink, and other names) was not initially developed to be a hot pluggable interface ("...and FireWire standards were first developed; then the ability to hot-plug-and-play was added."). A quick look at the web site for the 1394 FireWire trade association (www.1394ta.org) or their products showcase will quickly convince one that this feature is well supported in product fact, as well as in intent (www.1394showcase.com)
To suggest that USB somehow has a superior early history hot-swapping support is curious. If there were any such advantage, it was not readily apparent or meaningful. Among other things, a certain demo by Gates at a Comdex comes to mind.
The author also claims that relative to USE 2.0, FireWire cannot provide "Ease of installation," "hot plug and play," "cross platform compatibility," or "portability." Again, one wonders exactly what "special" definitions the author intended for these terms when the article was written. Would it not be fairer to say that real FireWire-based products have been providing this capability for some time, while USB 2.0 products merely hope to do so?
In terms of ease of installation, the author seems to be making the unfair comparison of the case of a PC with a built in USB 2.0 interface to that of a PC where FireWire must be added via a card installation. This might be a point if we were talking about USB 1.x, but we are not. Most customers have computers with neither USB 2.0 nor FireWire. So both must be added via a card installation.
If added, the potential problems and issues would appear to be the same. A number of FireWire adapters from multiple manufacturers are available, supporting Plug-and-play; so there is no difference there.
Considering that Apple, Compaq, NEC, and Sony have been shipping computers with built in IEEE 1394 FireWire interfaces for some time, it would appear that the author should have compared the case of having a PC with a built in FireWire interface to that of having to add a USE 2.0 interface. FireWire is now available built-in to the popular consumer iMac, adding even more to the installed base for FireWire.
As for "cross-platform compatibility," one again wonders what dictionary the author is using. Today, IEEE 1394 FireWire products are shipping for PCs and Macintoshes. Interfaces are available built-in from the manufacturers. What of USB 2.0?
Indeed, in their section on frequently asked developer questions, www.usb.org/developers/usb20/faq20.html, the USB Implementers Forum makes the following comment "The final specification will be published in Q1 2000. We expect systems and peripherals to be shipping by 2nd half of 2000." While FireWire-equipped systems and peripherals are shipping today, the spec for USB 2.0 isn't even shipping.
The author also failed to mention other issues, like support for isochronous, as well as asynchronous data. Isochronous support is important for digital video and digital audio. I understand that USB 1.1 lacks this important support. I don't know about USB 2.0. Does it? The author didn't say.
Due in part to its isochronous support, FireWire is beginning to be adopted as a consumer audio interface. What of USB 2.0?
Ironically, the author failed to mention that USB 2.0 is backward compatible with USB 1.0. This might be a huge plus for it, as computer manufacture's could begin replacing USB 1.0 with USB 2.0 devices. If USB 2.0 devices were to cost little, if any, more than current USB devices, there would be good reason to migrate to it from 1.x.
On the other hand, if compatible, one would want to understand the implication of mixing USB 1.0 and 2.0 devices off of the same interface. It would be a real minus for a USB 2.0 disk drive if it suddenly slowed down to 1.0 speeds because someone plugged in a USB 1.0 mouse. (More revenue sucking calls to tech support?) That alone would be a good reason to avoid USB for speed critical devices.
There is some information about USB 2.0 at the USB Implementers Forum (www.usb.org) Some of it is restricted to "members," although there is still a lot of useful information available to the public.
It appears the author's company makes USB products. If it were to fairly evaluate the merits of FireWire, perhaps it would be inclined to make products for that interface as well.
I don't doubt that USB 2.0 will turn out to be a great interface. I wish its manufacturers well in the marketplace. However, misinformation is not helpful.
Paul D. Cook
Faster Than A Speeding Bullet
I stand by my article ("USB 2.0: A Storage Bus Whose Time Has Come", page 32, October CTR) as written. Mr. Cook has some interesting opinions and observations, but we need to be on the same page for a meaningful discussion.
My article was not meant to bash other interfaces. It was meant to present USB 2.0's obvious advantages as a storage bus. Period. Although USB 2.0 is not available now, it will be soon. Engineers wanting to use USB 2.0 as a storage bus need to design it into their products now.
Mr. Cook might want to consider the following:
1. USB 1.x is successful in the market because of wide consumer acceptance. FireWire (1394) does not have this same acceptance, at least at this time. A consumer can go to any computer store and compare the large number of USB 1.x peripherals and the small number of FireWire peripherals that are available. Engineers need to design products for the market--if they want to sell those products.
2. Intel is committed to incorporating USB 2.0 into its South Bridge Chip, which means that literally all future PCs will support USB 2.0 in their chipsets. This practically guarantees the success of USE 2.0 and it will enable elegant handling of a mix of USB 1.x and 2.0 devices such as video cameras.
3. USB 2.0 has the transfer rate required for storage applications. In fact, since I wrote my article, the USB 2.0 specification maximum transfer rate has increased from 450Mbit/sec to 480Mbit/sec.
4. USB 2.0 is backward compatible with USB 1.x. (Contrary to Mr. Cook's letter, I did state this in my article.)
Finally, I would like to state for the record that, although we manufacture a USB 1.1 host controller, CMD Technology is known for its state-of-the-art RAID storage controllers. Many top OEMs incorporate our storage solutions into their products.
Although FireWire is well suited to some applications, our company sees more value in using USB 2.0 as a storage bus at this time. This is why we are encouraging. our OEM customers to design USB 2.0 into their products.
Believe me, we have not ruled out FireWire. If FireWire develops into a more promising interface than USB, we will jump on it faster than you can say "storage bus."
Product Marketing Manager
Storage Interconnect Products
CMD Technology. Inc.
|Printer friendly Cite/link Email Feedback|
|Publication:||Computer Technology Review|
|Article Type:||Letter to the Editor|
|Date:||Dec 1, 1999|
|Previous Article:||For DVD+RW, Another Delay; For DVD-RAM, A Surprising Design-Win.|
|Next Article:||New Technologies Emerging: Mark And Hal Discuss What's Next.|