USB

• 1. Introduction

  ›

  USB, the Universal Serial Bus, sits at the back of nearly every device you own. In 1996, when the first specification launched, a typical desktop computer bristled with a tangle of incompatible ports: serial ports, parallel ports, game ports, and Apple's own proprietary ADB connectors. Every peripheral demanded its own plug, its own settings, often its own manual. The USB standard was built to end all of that. What happened next reshaped not just computing but the way the entire electronics industry thinks about connectivity and power. The questions worth asking are: who built USB, and how did a connector standard become the nervous system of the modern world?

• 2. Seven Companies, One Problem

  ›

  In 1995, a group of seven companies sat down together with a shared grievance. Compaq, DEC, IBM, Intel, Microsoft, NEC, and Nortel were all tired of the same mess: a wall of mismatched connectors at the back of every PC, software that required manual configuration, and peripheral devices that could not simply be plugged in and trusted to work. Their goal was to replace all of it with a single, self-configuring interface that any device could use.

  The ideas they drew on were older than USB itself. The Atari SIO serial bus, designed in 1979 for the 8-bit Atari computers, pioneered the concept of a common peripheral bus. The Commodore bus, derived from IEEE-488 in 1980, extended the idea further. A separate consortium led by Apple, which included Sony, Panasonic, LG, Toshiba, Hitachi, and several others, had been developing the IEEE 1394 FireWire standard from 1986. USB's architects knew this landscape well.

  The engineer Ajay Bhatt and his team at Intel drove the technical work on the standard. The first integrated circuits supporting USB were produced by Intel in 1995, the same year the consortium formed. Joseph C. Decuir, who had worked on the original Atari SIO bus before moving to Commodore, joined the USB project at Microsoft and obtained one of the related US patents. When USB was later challenged by a patent troll, the Atari SIO bus was cited as prior art in its defense.

• 3. From Ports to Plug-and-Play

  ›

  USB 1.0, released in January 1996, specified two signaling rates: 1.5 Mbit/s for low-speed devices and 12 Mbit/s for full-speed devices. It was not immediately embraced. Few USB devices reached the market until USB 1.1 arrived in August 1998 and fixed the problems holding adoption back. USB 1.1 became the first version widely deployed, and Microsoft later designated the computers built around it the "Legacy-free PC."

  What made USB genuinely different from every interface that came before it was the experience of plugging something in. The standard is self-configuring: a user does not adjust speed settings, interrupt lines, memory addresses, or direct memory access channels. When a device is plugged into a running computer, it either works automatically using existing drivers or the system locates and installs the correct driver on its own. The USB specification also defines protocols for recovering from common errors, making the whole system more reliable than the serial and parallel ports it replaced.

  Small devices can draw power directly through a USB cable, eliminating separate power supplies. And because the USB logo may only appear on a product after compliance testing by the USB Implementers Forum, the presence of that logo carries a guarantee: the device has been verified to work as expected without configuration fights.

• 4. Speed, Generations, Naming Chaos

  ›

  USB 2.0, released in April 2000, raised the maximum signaling rate to 480 Mbit/s. That gave the standard room to support cameras, printers, and early flash drives without becoming a bottleneck. The USB 3.0 specification, released on the 12th of November 2008, introduced a fundamentally new architecture called SuperSpeed, running at 5 Gbit/s. Unlike USB 1.0 and 2.0, which were half-duplex and arbitrated by the host, SuperSpeed was full-duplex. It also introduced the USB Attached SCSI Protocol, which delivered generally faster transfer speeds than the older Bulk-Only-Transfer approach.

  USB 3.1, released in July 2013, doubled the maximum signaling rate to 10 Gbit/s using a new 128b/132b encoding scheme that cut line overhead to just 3 percent. USB 3.2, released in September 2017, added a second lane to reach 20 Gbit/s, but that two-lane mode requires a Full-Featured USB-C cable and remains uncommon. As of 2023, Intel had begun including the two-lane modes in its 11th-generation SoC processors, though Apple had not.

  USB4, whose Version 1.0 specification was released on the 29th of August 2019, is based on the Thunderbolt 3 protocol and reaches 40 Gbit/s. USB4 Version 2.0, announced on the 1st of September 2022, pushed that ceiling to 80 Gbit/s symmetric, or an asymmetric 120/40 Gbit/s for video-heavy workloads. It achieves this using PAM3 signaling, a technique that encodes three levels of signal amplitude rather than two.

  The naming history is a record of accumulated confusion. USB-IF introduced the SuperSpeed and SuperSpeed+ marketing labels, then retired them. Starting in 2022, the recommended naming convention became a simple statement of signaling rate: USB 5Gbps, USB 10Gbps, USB 20Gbps, USB 40Gbps, USB 80Gbps. Whether any given cable's packaging reflects the current scheme depends on when it was printed.

• 5. The Connector Keeps Changing

  ›

  Standard-A and Standard-B connectors, the two shapes most people recognized for years, were the original forms defined in the specification. The larger Standard format was intended for desktop equipment. Mini-USB connectors arrived for mobile devices, but they were quickly displaced by the thinner Micro-USB. The Micro-B connector became ubiquitous on smartphones and portable electronics for years.

  All of that is now giving way to USB-C, officially called USB Type-C. USB-C is reversible, meaning it can be inserted either way. It is the only connector required for USB4. It also supports Alternate Mode, which allows it to carry non-USB protocols: DisplayPort, HDMI, and Thunderbolt signals can all travel through the same USB-C cable. DisplayPort 2.0 over USB-C can support 8K resolution at 60 Hz with HDR10 color and can use up to 80 Gbit/s because it sends data in one direction only, using all eight data wires at once.

  Before USB-C, two host computers could not be connected to each other using a standard USB cable because USB enforces a strict host-to-device hierarchy. USB bridge cables existed as workarounds, and Microsoft popularized them through a utility called Windows Easy Transfer. USB-C and the dual-role device capabilities introduced in USB 3.1 changed this: two systems can now connect directly through a Type-C cable, provided both support role-switching.

• 6. Power: From Milliwatts to 240 Watts

  ›

  The original USB 1.0 and 1.1 specifications from 1996 and 1998 could deliver a maximum of 2.5 watts to a high-power device. That was enough for a keyboard or mouse, barely enough for a spinning hard drive. USB 3.0, released in 2008, raised the high-power ceiling to 4.5 watts. These figures defined the standard for years.

  USB Power Delivery changed the scale entirely. USB Power Delivery Revision 3.1, finalized in 2021, extended the maximum power delivery to 240 watts at 48 volts. This is the specification under which a USB-C cable can charge a laptop, power a monitor, or supply a device that would previously have required a dedicated power brick. The standard also defines Battery Charging modes, where a dedicated charging port signals its availability by placing a resistance of no more than 200 ohms across its data lines.

  A proprietary high-powered system called PoweredUSB, developed separately in the 1990s, reached higher voltages for use in point-of-sale terminals such as cash registers. It never became a general consumer standard, but it illustrates how early the need for more power from a USB-like connector was recognized.

• 7. Limits and Security Risks

  ›

  USB cable length is capped by the physics of the standard. USB 1.1 allows a maximum of 5 meters for full-speed operation and 3 meters for low-speed. USB 2.0 permits up to 5 meters for high-speed use. USB 3.0 does not specify a fixed maximum but requires cables to meet an electrical specification; for copper cable with AWG 26 wire, the practical ceiling is about 3 meters. The standard was designed for peripherals on the same tabletop, not across a room or between buildings.

  USB also enforces a strict tree topology with a master-to-slave protocol. A peripheral device cannot talk directly to another peripheral without going through the host. A host cannot broadcast to all connected devices at once; each must be addressed individually. Up to 127 devices can be connected to a single host controller.

  The plug-and-play design that makes USB so easy to use is also what makes it a security target. A device that looks like a flash drive can instead simulate a keyboard and type malicious commands. On a Windows computer, such a device can open PowerShell and download malware, an attack called BadUSB. A separate threat is the USB killer, a device that sends high-voltage pulses through the data lines to destroy or damage whatever it is connected to. Older versions of Windows would automatically run scripts from USB mass storage devices through AutoRun, and it is also possible to gain full system control by compromising a USB controller itself. The convenience and the vulnerability are two sides of the same design.