The incredible story of the origin of the USB port that changed everything
Ajay Bhatt was tormented with updating his computer, and in the end he saw the need for one connector, necessary to rule all
In ancient times, to connect anything to your computer — a mouse, printer, hard drive — you needed a zoo of cables. Perhaps you needed a PS / 2 connector or serial port, Apple Desktop Bus or DIN; possibly a parallel port, SCSI, or Firewire cable. If you have heard of these things, or if you have not heard, thank USB. When it was first released in 1996, its idea was directly in the title: universal serial bus . And to be universal, she had to be able to just work. “The technologies we are replacing — serial ports, parallel ports, mouse and keyboard ports — all required serious software support, and after each installation of the device, numerous reboots were required, and sometimes even opening the case,” saidAjay Bhatt , who retired in 2016 after working at Intel. "Our goal was this: I took the device, stuck it in the computer, and it worked."
It was Intel engineers in Oregon who managed to get the technology to work, and it was Intel who persuaded the industry to support this standard, since it was necessary to facilitate the work with PCs in order to sell them in large quantities. However, a person who was formerly a skeptic began to popularize the standard: to the surprise of many geeks, in 1998, under the leadership of Steve Jobs, Apple released a breakthrough first iMac that only supported USB. The increased USB 2.0 speed made it possible to use new peripherals, such as a flash drive that helped kill a floppy disk, Zip, and CD-R. Then a whole procession of connected devices followed: disco balls, massagers for the head, keys with passwords, an infinite number of charges for the phone. In the world now, according to one estimate, there are six billion USB devices.
Now on the heels of typical USB Type-A and Type-B ports comes the new cable design, Type-C, available on phones, tablets, computers and other devices - and, fortunately, unlike older USB cables, it is two-way. The next generation of USB4, which will be released this year, will be able to produce speeds of up to 40 Gb / s, 3000 times faster than the fastest version of the first USB. Bhat could not have imagined this when in the early 90s, as a young engineer at Intel, he tried to install a multimedia card. The rest is a story that journalist Joel Johnson recorded using several key characters. Their memories have been edited for clarity.
“I knew computers could be simplified.”
Ajay Bhatt : I think around 1992 - and in 1990 I joined Intel - I started looking at PCs. It always seemed to me that they were too complicated to use. This opinion was based on observing how hard my family managed computers and tried to accomplish some simple task, such as printing a document.
Even it was difficult for me to use them, but I am a technician. I had difficulty updating my PC when multimedia cards started to come out. I looked at this architecture and thought - you know what? There are better ways to work with computers, and this one is just too complicated.
Bala Kadambi(I / O Interface Architecture Manager, Intel Director of Technology I / O): If you go back to the origins of PCs, it was based on IBM hardware and documentation projects - and BIOS and interfaces. It was collected, believing that it would be used by specialists. But by the end of the 90s, it became clear that in the course of evolution, PCs should become easier to use.
Ab: The original goal was to attract a new class of users and promote new computer users. And it all started in 1992. I came to work and offered this idea to several managers, but I did not see much interest. People did not understand the benefits of something like USB, but I was worried about this idea. I knew that computers could be simplified, and that you didn’t need a specialist from the IT department to install a printer, configure a keyboard, mouse, or support multiple input devices.
BK : Both Intel and Microsoft imagined a business that went beyond the first 10 million users, and wanted the hardware and PC programs to be easier to use, to make them more convenient and standardized.
The first of these initiatives was the PCI bus (Peripheral Component Interconnect). It was conceived so that some systems in the PC case were easier to install, initialize, update and maintain. The initiative gradually evolved into the idea of plug and play [ plug and play]. PCI became the first standard 32-bit interface, and gradually evolved into PCI Express.
But even within the PCI framework, each type of peripheral had different characteristics in transferring data in and out from a PC. In some cases, it was required to use the adapter connectors outside the PC, or additional cards inside.
AB : The main difficulty at that time was the existence of many ways to connect to iron. That is, every time you changed something, it was necessary to carry out serious changes in the OS and in applications.
Everyone considered this the greatest difficulty, and my boss, in fact, announced that these changes should not be carried out: “I do not think that you will succeed. You don’t understand PC architecture. ” I replied: “No, no. We can fix it. Believe me. It is possible. ” It was hard for me to convince him.
BC: At that time, PC was not a laptop, it was a desktop computer. Laptops just appeared in that era, they were bulky, big boxes. Literally, I would say the device was portable if it had a pen. If there was no pen, it was a desktop computer.
By connecting these devices, you gradually drowned in a mess - your own standard for audio, your own for modem, your own for SCSI printer. Each of them had their own artifacts at work. Orientation of the connector. Can I connect on the fly, or do I need to restart the computer. Does the computer have software, or do I need to install it additionally from a floppy disk. Will everything work if you transfer the device from one interface to another.
Ab: I didn’t get any positive answers, so I moved horizontally to a parallel group in the company and started working under the supervision of Mr. Fred Pollock. At that time, the company employed very few " honored Intel researchers. " These are the most outstanding techies of the company. He was incredibly smart, one of the largest specialists in computer science. I talked to him, and he reacted like this: “I don’t know. You know what - try to convince yourself. " This is what I needed. I needed a man of sufficiently broad views that would allow me to take a chance.
But I did not rely solely on him. I began to convey this idea to other groups within Intel. I spoke with businessmen, spoke with technologists, and in the end I even talked with Microsoft. We talked with other people who eventually became our partners - Compaq, DEC, IBM, NEC, etc.
In fact, I had to not only draw up this idea within the company, but also enlist the support of external agents, and, of course, each person in each company had their own opinion on how everything should be. The general thing was that everyone eventually agreed that using a PC is too difficult, and even designing equipment for it is difficult. Something had to be done, and here it all started.
“A lot of assaults, a lot of discussion”
“In February 1992, when Intel and others were working on PCI and PnP, the group met in Redmond to discuss standardization of PC external interfaces.
BK : It was a spontaneous meeting. It was clear that even the formulation of our problem was poorly understood, not to mention how to solve it. At that moment, I already managed the team that developed PCI and PnP at Intel, so we had enough knowledge to understand how difficult this task would be.
We could not evaluate this in terms of the potential for future applications and the requirements of the technology. So the solution to the current problem was probably obvious, but in order to imagine what the PC interfaces would look like in a few years, we had to start research. Meet with companies. Talk to analysts. Chat with end users. Keep track of trends in business and the consumer market.
AB : Slowly but surely, I began to convince people from Intel of the need for all the subtleties of the requirements that we had, and which eventually turned into USB. I think that around 1993 we reached an internal agreement, and got down to work.
All these beliefs took away about a year and a half. There were many assaults, many discussions in the early stages of development, when we needed to overcome people's skepticism and form the same idea about how to solve the problem.
By the end of 93 or the beginning of 94, I assembled a small team. At Intel we had internal working groups that generated ideas, engaged in analytics and wrote specifications. And we regularly worked with external partners.
Then the project was called Serial Box. He had no name. Everything came from technology.
Work on a Serial Box in a diner at 2 nights
Ajay and Bale, who met through collaborative work on the PnP initiative, were joined by Jim Pappas, an expert on I / O systems, and marketing experts. In the summer of 1994, they formed a working group.
Jim Pappas (CTO, now Director of Technical Initiatives at Intel): We were a very focused and dedicated team. Our four - I, Bala, Ajay and Steve Wally, were very, very active and very closely interacted - especially Bala and I. I have not called him home for about fifteen years, but when I picked up the phone, my fingers dialed his number ourselves - so often we talked. That was incredible.
We arranged what was called a charged breakfast. Suppose we had an appointment with a company in their territory. We could fly into place at different times from different cities. At one or two nights we met at some diner, Denny's or the other. And there we arranged what was called a charged breakfast. “What will we say tomorrow, what do we need?” We had the habit of not only working together, but also working almost around the clock, and it was an incredibly interesting experience.
BC: We literally went on a trip for a year. We visited about 50 companies across the entire spectrum of the industry - printers, scanners, communication devices, industrial controllers, keyboards, mice, joysticks, modems, etc. I am shocked at how much interest there was in such a simple thing as a standard external interface. Enthusiasm was born because all these companies had their own requirements, and they believed that their market opportunities were limited by current interfaces.
However, the biggest problem was that we did not solve special problems, only eased the current state of affairs. Everything could already be stuck somewhere, so the difficult question arose: “why do we need to accept the standard before the next release?”
The team of Ajay, Bala and Jim has grown into a larger group at Intel. It was divided into separate disciplines: protocols, the organization of bits, electromechanical issues - connectors and cables - a business group, an adaptation group.
AB : We organized to carry out an attack on all aspects of the technology and successfully enter the market, because we wanted to take into account all the nuances. We wanted to not only release specifications, but also help developers release technology-related products.
We did not stop after we finished with the specifications, and came up with a recipe for the development of various proposals. We also did the so-called “Interoperability programs” - when different manufacturers came together, they were tested on pre-prepared tests, and we were convinced that everyone followed the specifications and that all devices would work together without problems.
And although we were an alliance, we were like a startup, paying attention to all aspects of not only specifications, but also product development, and, as a result, its introduction to the market.
DpAjay led the technical specifications, Ball engineers, I led the entire program. We assembled a group of engineers. In an Intellectual situation, it was called “two in one box” when you have two managers. I asked Baloo to join me in managing the program. We stood back to back. He looked inside Intel and managed the engineers, I followed the progress in the industry.
BC: Jim was engaged in external relations, external building of the industry, assembled a group of promoters, an industry forum. He has a talent for introducing technology aspects to the market. He often delegated me speeches, but we worked on them together. At the end of each day, we exchanged information, keeping each other informed. They called at 11 pm. This went on for years. The phone rang, and our spouses thought: “This is probably Bala. This is probably Jim. ” That was the regime.
Ab: I think we succeeded because everything that we did was well implemented. And I knew that this would be so, since we had multifaceted teams, in which software and OS experts worked. We had people who knew how to create systems, for example, IBM and Compaq. We had people who knew how chips were made, for example, Intel and NEC. We had a Nortel company that knew how to create telephony and other things that eventually became very important. By putting together a team of experts, we were able to reduce the risk, and guarantee specifications of wide application suitable for various options.
BCAjay was the heart of specs. That was his passion. He was also passionate about understanding the requirements to ensure that the specifications met the requirements. Engineer Jeff Morris then moved from Santa Clara to Oregon to work with my team. He knew that he really wanted to work on this project. He wrote a good half of the first specifications along with all the technical descriptions needed to develop the technology.
AB : We finished the specifications around 1995. The trade fair COMDEX was held. Our goal was to finish the specifications in time for it, by November 1995. Then we started working on products and all that. It was a long way, but in the end, the industry realized that this work really solved many PC problems.
Problem with Firewire and other interfaces
Computer companies really wanted to facilitate the connection of devices, but among applications that required an increase in the speed of interfaces, video stood out especially. Digital multimedia was still in its infancy, but the transfer of video to computers and downloading it was one of the main tasks of manufacturers of computers and peripherals. Intel engineers, working on an interface that will later become USB, also explored possible faster alternatives.
BC: In general, streaming multimedia, transferring video from and to a PC was an area that we could highlight and say: “You need to be able to do something like that.” I think that in general, most companies recognized that if the device is easier to use, then it will be easier to purchase, update, support and service. They will also receive fewer returns and less support calls.
Gathering requirements, we simultaneously evaluated technologies that could fit into them. We clearly did not want to invent something new if there was already something similar enough or good enough.
We studied about 12 different technologies. The most obvious was IEEE 1394, which was later dubbed Firewire. When I first started appearing at these commission meetings to check if the technology worked, 1394 was a 10 MB interface. It seemed like a technology that was looking for a problem that it could solve. They already had something, but they were not sure what to use it for, and this technology was developing. It was a little more complex and expensive than was necessary for a PC. On the other hand, she had potentially useful elements.
We looked at the then generation of technologies such as Ethernet. We studied the interfaces for audio. Apple then had a GeoPort interface. We even talked with Apple to find out if they would be interested in developing this project. But it did not work out. Another industry standard was Access Bus.
Ab: I personally traveled to a huge number of different forums. I talked to people from related fields, and said: "Guys, let's collect all our applications." There is a MIDI interface for music, and it is used by many synthesizers, keyboards, and all that. I remember how I had a meeting with key telephony manufacturers in Dallas, because there were a large number of external partners. We tried to convince people that computer telephony can be done using something like USB, and many people thought that we could not support certain things of this kind. It was believed that HP wants to make printers communicate with a computer via infrared.
BC: I would say that in parallel for a couple of years USB, 1394 and Access Bus have been developing. In the years 93-94. And then broad public support for USB was formed. Then it was called Serial Bus. We haven't come up with a name yet.
How USB got its name
BK : It took a lot of committee effort to name the USB. We had three ways. Someone thought that the name, consisting of numbers, did not take root. Too tech. You do not need to call it something like 1394. This is the specification number, but you need to do something that users can associate with. We tried to come up with consumer names. And then we decided that we were too far removed from what the USB project was.
If you haven’t noticed, Intel likes acronyms. If we turn to our organization, then many names of teams, organizations, projects, technologies consist of acronyms. We had to build on this, as well as on the universality of the solution. We played around with this. How can we expand this?
On the other hand, the word "tire" seemed counterintuitive, but the industry knew what it meant. Therefore, we left him. Other interfaces were parallel - SCSI, parallel port, etc. Our new standard was fashionable. Economically simple. I wanted to put these elements in the name.
Therefore, the universal serial bus option pulled ahead. The idea was to use the word bus - not only the "bus", but also the "bus" / approx. transl.], as something that delivers you from one point to another, efficiently and reliably.
In the end, I think it won due to versatility. That is what we tried to do.
Dp: COMDEX was the largest show in Las Vegas, and in 1998 we rented a large hall, we had a large showcase. We rented a large pavilion, where we held a press conference, we connected 127 devices to a PC and hired Bill Nye [a famous popularizer of science / approx. trans.], which connected the last device to show how much this single port could support on a PC - we had a full scene of different printers there! We walked, shook with one mouse, then another, or typed here and there.
Well, why is this connector not double-sided?
AB : Good question. We thought about it, but our goal was to make a very cheap port, and at that time we tried to solve all USB problems with two wires. At that time, if we added wires to make the standard plug-in in any way, we would have to add a bunch of wires and a bunch of silicon. Wires and contact cost money, so we decided to stay on the cheapest option. Serial ports and parallel ports had options with 25 pins, 36 pins, and so on. The cables were thick and expensive. We tried to solve all the problems. We were for fewer wires. Looking back, I can say that a two-way connector would be better.
Ab: Our goal was to say that this should be an interface that will work with a mouse, with an expensive printer or with a digital camera. We looked at a range of products. On the one hand, we wanted it to be simple enough so that the cost was very low. On the other hand, we wanted to scale, and today, when we talk, USB works at tens of gigabits / s. And the first worked at 12 Mbps. We have come a long way to scale.
Microsoft calls Betsy Tanner to save USB
DP : One of the people we met at Microsoft was Betsy Tanner, and at that time she was the chief mouse engineer. I talked to Betsy and said, “if the day comes when you decide not to use USB for the next mouse from Microsoft, I need to know about this.” And she says, “okay, reasonable request.”
We developed USB - at first it was supposed to be a 5 Mbps bus, which at that time was faster than anything from the back of the PC case. By today's standards, this is not fast, but then it seemed fast. And we needed a great speed in order to be able to branch connections through the hubs, and, in fact, all devices connected to the same port would share this speed among themselves - not necessarily they would be used simultaneously, but we wanted to so that everything is reliable enough. One day Betsy called me and said: “Jim, you asked me to call if we decide not to use USB for our mouse. I’m calling to say - we can’t do this because we have a problem. ”
I asked: “What is the problem?” She said: “Well, 5 megabits is too fast for us.”
I said: “This bandwidth is not required for a mouse, and secondly, I’m worried if we can meet the specifications for electromagnetic interference. Signals traveling through the wire become an antenna. Will I have too much EM radiation to create digital noise? "
She said," We could solve this problem through cable insulation, but it adds 4 cents per foot to the cost. On a six-foot cable [1.8 m] you get 24 extra cents. I can not do it. Secondly, how can I insulate the cable if the mouse needs a simple cable. He should not influence her movements, and I am afraid that if isolation is made, it will become too hard. ”
I said: "Betsy, what would suit you?" She said: "It would be convenient for us to work with two megabits per second."
I said, “Damn, this is very slow. Can you give me a week? ”
She agreed. I turned to the team, we discussed the Microsoft problem, and then we came up with the separation of high-speed and low-speed modes. We brought high speed to 12 Mbps. We limited the low speed to 1.5 Mbps, 3/4 of the maximum required for the mouse.
We saved Microsoft, we saved the mouse. And I think that this call from Betsy saved the program. One of the reasons for the success of USB is that the port has met cost limits. It did not cause a significant increase in the cost of PCs. You can even say that over time, he reduced this cost.
Apple is completely not interested in working with us
In the confederation of companies that helped bring the USB standard to life, one notable player was missing. But in 1998, with the release of iMac, Apple was the first to make USB the only interface on their computers. It was Apple, not Intel, that became the first well-known computer company associated with USB.
AB : This is interesting. Apple was not on the list and they had a competitive product, 1394, or Firewire. Apple had its own interface. They were still famous for their ease of use. But at the end of the development of specifications, it was Apple who came out with the first product. The Windows-based system was moving from DOS to Windows and from Windows 3.1 to Windows 98.
Remember, we were not marketers. Our idea was to significantly change the computer industry. That was my motivation as a specialist in computer science. I wanted to eliminate the clumsy interfaces, as they limited internal extensions, and limited some computer applications.
And when we started all this, we talked with Apple, and they were completely not interested in working with us, they wanted to go the other way. When they accepted the specifications, we knew that we had done everything right and started to solve the problem we needed. We were only happy. We believed that this pie should be increased, and then everyone will get a significant piece. We were not disappointed. We were delighted, and every time something new came out, we were even happier, and this confirmed our idea that we were solving the right problem.
Intel ad with Ajay Bhatta: “Our rock stars are not like yours”
Everything was with USB
DP : Since the fall of 1996, USB ports have begun to appear on PCs. In the fall, Microsoft released Windows OSR 2.1, if I'm not mistaken. She supported USB. But it had to be installed; OEM could not sell it with new machines. Peripherals came out, but it was not like what happened in 1998.
When Windows 98 came out, there was just a dam break. USB devices have flooded the world. I remember how Steve Wally and I were in Tokyo. We went to Akihabara, an electronics sales area in Tokyo, and went to one of the major electronics stores. We started walking and looking for something with USB. This was before the release of Windows 98 or around the time. We walked, did not see anything special. Someone came up to us and asked if we needed help, and we said, “yes, we are looking for USB devices,” and he says “ah, it's on the fifth floor. The entire fifth floor is dedicated to USB devices! ”
In this electronic supermarket, the whole floor was dedicated to USB devices. It was an amazing moment. You come in, and there are whole rows. And everything was with USB.
BC: Who would have thought that the connector we described in the early 90s will still be used? This rarely happens. We had limitations in terms of cost and speed. It was designed for a desktop computer, not for smartphones. Looking back, I consider it a miracle that we were able to do everything we did, and that this work stood the test of time - that we were able to develop on its basis, improve power, speed, everything that was done in USB2 and USB3.
USB miniaturization has taken us beyond the PC era and right into the mobile era. We have developed other protocols that have appeared since then, except for USB. And all the good things we put into Type-C.
Ajay Bhatt and Bala Kadambi at the 2013 DP Inventor Award
: With USB-C, you can charge your laptop using one USB port. You do not even need to make a separate port for power. It was a serious matter. Who would have thought?
BC : Defining a new connector is always a transition issue. We very carefully approached this issue. It took us six years to create Type-C capabilities. Behind this is the enormous industry work.
There are two aspects to this standard. The first is the interface you stick in, and it changes. And the interface behind it, between device drivers and the OS, between the device and drivers - this does not change. It worked great for 20 years, and goes into USB-C.
Dp: Bala once made a great slide. Sean Maloney was one of the senior vice presidents, and Bala somehow photographed Sean's computer from behind. And there was just a rat nest of wires. It was a visual representation of the industry at that time, and this confusion conveyed this state.
And he made the second slide with USB - the first one hundred million devices. I remember how we inserted this slide into one of my presentations, and the audience laughed. And then, after a few years, we already produced 2.2 billion devices per year. And no one laughed. The success of this thing was phenomenal. She has become a ubiquitous connector.