- A TechNote on Infrastructure and Performance
- Steven Taylor, Editor-in Chief, Webtorials
Size matters. But functionality and durability matter too.
Despite all of the attention over the years devoted to medium-high and high layer protocols such as ATM, frame relay, and TCP/IP, none of it works without a way to pass signals. Information (data) signals are passed across a physical interface (wired, optical, or wireless connection) while electrical power is passed on via wired connectivity (with rare exceptions).
"In the beginning," when many of us entered the data communications/telecommunications industry, the default interface for data transmission was the RS-232 specification, which specified a 25-pin connector dubbed a "DB-25." In fact, any data communications installation had rows upon rows of equipment with many, many, many DB-25 connectors on the back. In the case of modems, the RS-232 interface was the standard connection between modems and other equipment. In other cases, "terminals" (and later PCs) would be connected to the network switches and multiplexers using DB-25 connectors.
A DB-25 connector had three primary characteristics - size, function, and durability. And the size was a major drawback. A connector measures one-half inch by two inches, so each connection took a full square inch of space in a rack (with NO space between connectors. In fact, this became a severe limiting factor in the density of connections for a lot of equipment, so various schemes were devised to try to economize on this.
Here's a great example of the newest equipment from the 1980s, a Timeplex "switching statistical multiplexer." I helped with beta testing for this product, and I can testify that this is about as dense as you can pack DB-25 connectors, especially considering the nightmare of cable management. Here's the source for the photo, which was found in an old advertisement.
The DB-25 was over-engineered, as was most equipment in the 1970s and 1980s. As noted by its name, a DB-25 had 25 pins for various signaling functions. Seems like a lot for an interface used to "transmit" and "receive" data! The minimum number of connections that were critical were pins number 2 (transmit), 3 (receive) and 7 (ground).) But all 25 had assigned functions and were sometimes needed, like the "clocking" signals for synchronous data transfer.
And the connectors were DURABLE. It was really tough to damage a connector, and, in the rare event that one could bend a pin, it could usually be repaired with a pair of needle-nose pliers. Further, two screws were used to SECURE the connectors. So, while it was possible to have a bad connection, it took a LOT of effort.
Fast-forward thirty years.
We've gone through a plethora of physical connectors. Coaxial Ethernet. Optical interfaces. Ethernet over twisted pairs. V.35. Firewire. Countless proprietary versions. And, of course, three generations of Universal Serial Bus . And the USB comes in three sizes - standard, mini, and micro.
I have no problems with standard sized USB connectors, and very few with mini-USB. It's fast, and it provides connectivity for both data transmission and electrical power. And micro USB is, in theory, an answer to lots of prayers. It's tiny (and deserves it's "micro" name). And it provides a "universal" interface that can be used for everything from tablets to cell phones to computers to video cameras to... you name it. And equipment manufacturers have almost universally adopted it.
But the problem is that it is so flimsy that it is a very common point of failure for physical reasons. And even though you won't see many micro-USB connections in the data center, it's a macro-headache for supporting both corporate and consumer devices. As the BYOD trend escalates, the corporate IT staff will be faced with supporting these devices, and micro-USB connectors are not - and never will be - ready to assume the duty cycle that they need to sustain.
Compare the micro-USB with the ancient DB-25. That connector was physically snug, could be secured with a screw at either end, and was seldom moved. By contrast, a typical micro-USB connector is connected and disconnected several times a day, it has no reliable method for securing it (other than friction), and it relies on a hope that there's a good physical contact.
It's this last point that is the true Achilles heel. The micro-USB connector is designed such that the connector on the cord is (hopefully) sufficiently springy to make a good connection and the data connections are molded into the connector.
The "equipment side" of the connector is even more vulnerable to failure. It is, at best, an extension of a circuit board. Easy to manufacture, but extremely fragile. So you go directly to a trace on the circuit board, and you hope that the two connectors make sufficiently good contact.
Here's a photo of a micro-USB connector (left) and a micro-HDMI. You can probably also see both even better if you look at your mobile device. The micro USB depends on the hope that two extremely thin and fragile pieces of circuit board will touch each other. At least a micro-HDMI has real pins and sockets.
Personally, I've had three devices "fail" in a period of less than two years because of failed USB connections. Repair shops confirm that the vast majority of cell phone and tablet repairs are due to bad micro-USB connectors - right up there with broken screens. And the connections are so flimsy that an enthusiastic user can even force the connectors together upside down. (Yes, this invariably results in a hard failure.)
What's the solution? I don't know. Maybe we have to go back to larger form factors simply to have a more durable connector. After all, a powered Ethernet connector isn't all that huge. And WiFi connections (except for power) certainly offer a great physical layer.
The bottom line, though, in my humble opinion, is that this is a problem that will get worse before it gets better, and forward-thinking IT shops need to have a plan for a profusion of devices that use this Unmitigated Stupid Blunder
I would be surprised if this trend were to change, as it allows manufacturer to sell more devices.