Some thoughts on the USB-IF and USB certification

A rant about an industry standards organization

A couple of weeks ago, yours truly wrote a news story relating to Texas Instruments (TXN) gaining USB-IF certification for two of its upcoming USB 3.0 host controllers that reflected on the fact that none of its smaller competitors have achieved certification, yet all of them are already shipping silicon. We did contact the USB-IF, and received a reply curiously signed “USB-IF administration”.

We responded to that, but did not receive a reply to it or subsequent emails. It’s very possible that the USB-IF didn’t take us seriously enough, or that we somehow managed to upset them. Either way, we need to clarify one mistake in the Texas Instruments story, TI isn’t a part of the board of directors at the USB-IF, although one of our sources stated that “TI is a very substantial financial contributor to the USB-IF”, whatever that implies. It’s also important to point out that TI is one of the companies that helped develop the USB 3.0 standard alongside HP, Intel, Microsoft, Renesas and ST-Ericsson.

The USB-IF told us that it doesn’t receive any form of funding from any of its board members or member companies, instead all of its members have to pay an annual membership fee which is set at $4000. For that small fee its members can apply for a unique vendor ID and have the right to put the USB logo on their products after they pass certification. On top of this, member benefits include participation in various developer conferences, compliance workshops, device working groups and marketing events. A company can also become a USB-IF non-member licensee for $2000 which allows it to use the USB-IF logo and gain a vendor ID, although this fee is good for two years membership. This all seems very fair and we don’t have any criticism with regards to this, but it’s not quite the full story.

As far as we understand it, a company can sponsor the USB-IF at different levels. Being one of the USB-IF gold sponsors at an event brings with it additional, if costly perks, such as being promoted by the USB-IF more than non-gold sponsors. This is not what we’d call a big deal either, as it’s pretty standard in the industry, for better or worse and anyone willing to pay up can be a gold sponsor.

However, what surprised us was that one of our sources told us that many companies in the industry have held off going with various USB 3.0 host controller implementations on the back of TI promising its customers that it was going to have its USB 3.0 host controller ready by, well, by quite some time ago. Although this didn’t happen when promised, the company has as yet to ship any kind of quantity for its two host controllers, due to its promises, many companies have held off using alternatives. The other important reason for this is that the only USB 3.0 certified host controllers currently shipping come from Renesas and Fresco Logic, although in the latter case we’re talking about a single port implementation intended mostly for notebooks.

This would seem to imply that TI was certain it would pass certification way ahead of time. We are by no means saying that TI has done a poor job with its host controllers, nor that anything illicit has taken place, as we have no proof of that, however, it seems a little too co-incidental.  What also surprised us was that we were told by the USB-IF that “the USB-IF Host certification is based upon a first come, first served basis“. So does this mean that TI got in before ASMedia, Etron, Fresco Logic and VLI? Well, we don’t know, but considering that the other four companies all have working silicon, some of them even in shipping retail products, we have a hard time believing this to be the case.

The USB-IF is meant to be a standards organization that is supposed to help the development of USB products and further the eco-system, but in the case of USB 3.0 it almost seems like the USB-IF is holding at least some companies back. As stated before, it’s very possible that we have four companies here that just aren’t able to meet the requirements by the USB-IF, but we’re fairly certain that this isn’t the case at this point in time.

The reply from the USB-IF goes on to say “The USB-IF does not base its certification on whether manufacturers’ products are shipping, in stock or being deployed/adopted by OEMs/ODMs; many companies get their products certified months in advance of announcing or shipping products“. This is fair enough, but how can the companies in question get certified when there’s some very obvious roadblocks in their way?  We’ve seen working samples of USB 3.0 host controllers from both Etron and VLI some nine months ago now, yet somehow these products are still lacking certification.

We’d also like to point out a small piece of information here with regards to the USB-IF’s testing facilities. As you’re most likely aware, for a product to carry the USB logo it has to have passed certification and this certification process is performed in a lab. The USB-IF claims that these labs are independent, but of course qualified by the USB-IF. However, it turns out that not all of these labs are as independent as you’d think, as at least one of these testing labs have very close ties with Intel and we’ve been informed by one of our sources that AMD was none too happy about handing over its upcoming chipsets with USB 3.0 support to Intel for certification testing.

On top of all of this, the USB-IF hasn’t even finalized the certification procedure for many classes of USB 3.0 devices, and we know for certain the process for testing USB 3.0 hubs has as yet not been finalized. Again, this is something you would have expected to have been decided upon a long time ago, but apparently not. It almost feels a bit haphazard by a standards organization to not have worked out ways of testing common ecosystem building blocks of a standard that they’re in charge of.

The USB-IF also told us that there “are 200 certified SuperSpeed USB products that range from host and device silicon, notebooks, add-in cards, storage devices, media player and flash drives“. Oddly enough, the USB-IF data base only lists 167 certified products, of which many are pure IP implementations that may never actually end up in a product.

Part of the problem also seems to be that USB isn’t a hot topic, most of us take the USB interface for granted and could care less about what goes on behind the scene as long as we’re getting devices that just plug in to our computers and work. That said, it’s now possible that we’ll end up with devices in the market that don’t work as intended due to the time it takes to get USB 3.0 implementations certified. The reason for this is simply that companies that are investing millions of dollars into developing new products grow tired of waiting to get a stamp of approval so they can sell their products.

Does the USB logo really matter? Well, maybe, as it does at least tell you that the product for all intents and purposes was tested, unless of course the company that made the product just stuck the logo on without approval and didn’t care about having it certified. In fact, the USB-IF issued a statement a while ago about the usage of USB ports for other intents than for USB signaling and said this was a big no-no. We have no doubt that there are plenty of USB cables and peripherals on the market that never passed certification, but as far as consumers are concerned, as long as it works, that’s all that matters.

Standards are a good thing, don’t get us wrong, we’re all for standards, but a ubiquitous standard like USB should be more transparent, not one that is a closely guarded secret. The organization appears to be more concerned with politics rather than enabling the ecosystem that it was put in place to build and grow. There’s no reason as to why USB should be a closely guarded secret that only paying members have access to, as whichever way you look at it, computers wouldn’t be what they are today without the USB standard. It’s highly unlikely that the USB standard will ever become open and free , but that doesn’t seem to bother many companies. Unfortunately, many organisations don’t learn from past mistakes and tend to get tied in with more closed standards as time progresses.S|A

Update: Added don’t to the last sentence.

The following two tabs change content below.