Connect and Extend
Even in the earliest days of computing, it was obvious that a computer had to have ways for information to enter and leave the system to connect to other hardware. Devices external to computers that added functionality were peripheral to the main system, and that’s why we call them peripherals to this day. A computer could also connect to some kind of communication system, like a network or hub.
We use a port to connect to peripherals, networks, and other kinds of external systems. A port can connect two computing devices, too, like two smartphones or a tablet to a computer. The connection requires cabling—typically with metal inside its insulated cover, but sometimes containing a fiber optic strand—to connect a central computational thing to something else that provides something that isn’t built into the main unit.
Discover Peripherals and Devices
Peripherals are historically divided up among input, output, and storage, along with those that can do both, such as a multifunction printer. Other devices you can connect to include an Ethernet switch for local area networking, to a smartphone (or from a smartphone), and to power adapters.
Here are some examples among even narrower categories:
Data input and pointing: Keyboards, mice, trackpads, and trackballs, game controllers, microscopes
Pixel or rasterized output: Displays, printers, laser cutters, platesetters, 3D printers, and magnifiers for the visually impaired
Audio: Mics, headphones, headsets, earbuds, and amplified or line-level sound output
Image acquisition: Cameras, scanners, microscopes
Storage: Hard disk drives, SSDs, read-only and read/write CD, DVD, and Blu-ray
Terminals: TTYs for the hearing impaired, Braille keyboards/readers
Mobile computers: Smartphones, tablets, gaming systems
Hubs and power adapters: A hub is a collection of ports, optionally providing power. Older power adapters acted more like passive conduits for AC-to-DC conversion, but modern ones feature tiny computers and may have hub features built in. The hub and power adapter negotiate acceptable standards and the levels of power that may be passed.
Learn About Ports
To connect to peripherals and other devices you need ports: these are typically receptacles into which a plug slides, snaps, rotates, or magnetically connects. You need to know a little jargon that’s commonly used before we proceed as you’ll find printed above ports, on cables, in manuals, and in online instructions.
Here’s what you need to know:
A host, often but not always a desktop or laptop computer, coordinates activity over a port. A host may have more powerful technology inside it than anything it’s connected to, or it may connect to a similar device in host-to-host connections (Figure 1).
Figure 1: A host computer contains a primary circuit board, or logic board, that holds the CPU, controllers, and connected ports. A cable connects a port to a peripheral, like a printer. (Background image: Umberto via Unsplash) Ports are connected internally to a controller, a specialized module that’s part of the logic board (or motherboard) or plugged in as an expansion card (see An Internal Bus Detour). A controller handles one or more interface standards. Hosts, peripherals, networking hardware, hubs, and docks also include controllers.
A controller operates a bus, a general description for any kind of information flow among devices.
A peripheral might be powered by the port (bus powered) or plugged into external power (self powered). Conversely, a peripheral can be plugged into AC power and pass a charge to a connected computer or other device. (This is most typical with a display, dock, or power adapter.) Audio, video, and networking connections typically pass very low power.
A port may pass data, power, or both. Data or power may pass only from the host to an attached device or in both directions.
Depending on the port, it may require additional authentication before it can be used. This typically occurs in software at an operating system level rather than with hardware components.
A desktop or laptop computer traditionally had many specialized ports, often one for each kind of major purpose, with only a single multi-purpose bus standard, like serial, USB, or Thunderbolt (Figure 2).
Every port shares a common set of characteristics:
Something is plugged into it to pass data or power (or both) between a host and peripheral, a host and a power adapter, or two hosts, which may be computers or other devices.
Electrical signals of various voltages are passed over unique wiring paths. Each path may have a unique purpose in the standard. These used to be round or square pins or spring-like wires; most modern connectors use something more akin to flat metal wiring that uses friction and pressure to complete a circuit. (As a modern user, you never need to worry about the electrical signaling details!)
The thing plugged in has to be complementary, designed to fit into the slot; sometimes that’s not as obvious as it seems. In other cases, jacks and plugs are designed with backwards compatibility, so an older standard can fit into a receptacle designed for a newer one.
At one time, it was self-evident what standard went with what port: each port had a unique connector, as I describe next in Look Backwards at Ports, designed for a unique purpose. With the introduction of a common jack/plug type, that’s become more complicated, explained in USB, FireWire, and Thunderbolt.
An Internal Bus Detour
Computers include internal buses that once were able to push data many times faster than any external port. These buses, which are connected directly to a host’s logic board, originally relied on broad cables with many wires that allowed high throughput—more wires meant more simultaneous data.
One or more buses connect internal hard drives and SSDs, as well as internally mounted drives with removable media, if you still use such a thing. In the old days, that might be one of the Zip, Bernoulli, and SyQuest formats; more recently, optical drives for CD, DVD, and Blu-ray (or all three).
A second kind of bus appeared largely in office, professional, and hobbyist desktop computers that could be opened up. Card slots were wired directly into the main logic board with a similar broad connection. An expansion card became literally an expansion of the logic board, again using a lot of wires (this time, metal circuit traces, more or less flat wiring). You plugged expansion cards into card slots. This would let you add high-performance ports for video via a GPU (graphical processing unit) card, onboard coprocessing for faster computation, networking, and other specialized purposes. (Apple and other companies even offered cards for Macs with a Windows PC on them.)
You can still buy high-end computers with internal bays and card slots, but due to improvements in data standards and cabling, externally connected devices can often achieve the same performance as if they were plugged into an internal bus. For instance, you can purchase an external graphics card, an eGPU, that connects via Thunderbolt to computers that lack internal slots (or ones big enough) to get the same benefit as if it were plugged into the internal bus. Likewise, the fastest current SSDs read and write data just as fast via Thunderbolt as inside a computer.
Most computers you purchase today rely on PCIe (Peripheral Component Interconnect Express), an evolution of earlier internal bus standards. It appears to be the winner in the consumer and general professional market for both internal and port-based connections. Thunderbolt’s data transmission standard directly relies on PCIe protocols.
I explained earlier that a controller is the hardware module that manages a bus by negotiating data rates and transferring data from a host or peripheral on and off the bus. For greater throughput, many computers include multiple controllers, each with one or more of its own buses. Each bus can operate at the full data rate available from the controller. If your computer has two Thunderbolt controllers, you may be theoretically able to move 80 Gbps in and 80 Gbps out across multiple ports at the same time.
Look Backwards at Ports
During the broad expansion of personal computing in the 1980s and 1990s, manufacturers often sold hardware that had a set of built-in ports and most users found little reason for expansion:
In 1990, a Windows-compatible PC might include:
Two round PS/2 ports for mice and keyboards
A trapezoidal VGA port for an external display
One to three 3.5 mm jacks for audio
An RJ-11 jack for connecting a phone to an internal modem
Serial and parallel ports for printers and external drives
Networking would commonly be added only for office environments through a card that had a coaxial terminal (10Base-2, cable TV-like wiring) or RJ-45 jack (10Base-T, phone-like wiring) for Ethernet.
A 1990 Mac, like the Macintosh IIfx, would include:
The round ADB (Apple Desktop Bus) for keyboards and mice
Round serial ports for modems, printers, and AppleTalk, a proprietary networking standard
A rounded rectangle 25-pin SCSI connector for drives
A 3.5 mm headphone jack
VGA output (through a separate video card)
While I’m taking a snapshot of circa 1990, you could find a similar set of ports among computers for several years before and after. A next wave was coming that helped slim down the port profile.
USB, FireWire, and Thunderbolt
After a period of roughly 20 years of port profusion, the many different kinds of connections necessary became simplified. For a common set of peripherals, like a keyboard, mouse/trackball, external optical drive, external storage drive, printers, and scanners, you might be able to rely on one or two different port types. A series of three new standards replaced almost everything except, initially, video and Ethernet ports:
USB 1.0 arrived in 1996 for low-throughput input devices like keyboards. In 2001, the release of USB 2.0 allowed speeds high enough for external storage, scanners, and printers. USB 3.0 followed in 2008 delivering even faster performance; USB 4.0 is the latest. (See USB and Thunderbolt for more.)
Apple FireWire (also known as IEEE 1394, Sony i.Link, and Texas Instruments Lynx) appeared in 1997 with performance high enough for external storage as well. The standard was broadly used by Apple and lasted under two decades. (For more on FireWire, see Appendix A: FireWire.)
Thunderbolt first appeared in 2011 with staggeringly better performance than USB 2.0 and FireWire. It has iterated through three generations since. Thunderbolt was the first standard to incorporate video, data, power, and networking over the same port and cable; USB followed. The latest version is Thunderbolt 4. Its successor, not yet numbered—but probably Thunderbolt 5—was announced in October 2022. (See USB and Thunderbolt for the full details.)
USB, FireWire, and Thunderbolt all used different connectors. The demise of FireWire meant one less to manage. Dramatic increases in throughput for USB and the introduction of Thunderbolt took the pressure off using internal or specialized connections for hard and optical drives and, increasingly, SSDs. That led most computer makers to drop internal expansion slots and external storage-specific connections outside of expensive professional models and systems designed for gamers, hobbyists, and data centers to assemble from scratch.
Finally, with the introduction of Thunderbolt 3, following USB 3.1, USB and Thunderbolt converged on a single connection type: USB-C (Figure 3). This compact, flat connector with rounded edges can be inserted lengthwise in either of two 180° orientations—it’s “flippable”—unlike any other USB formats or Thunderbolt/Thunderbolt 2 connectors.
This commonality is great, but it has led to the situation where looking at a jack, a plug, a cable, or an adapter doesn’t immediately reveal the level of compatibility and maximum throughput. I get into the trouble that can cause in USB and Thunderbolt, and how to work around it.
USB-C has built-in flexibility that allows many different kinds of data streams to pass over it. This resulted in many computer makers picking USB-C for data, video, and, for laptops and mobile, power; some desktops and laptops also retain HDMI or DisplayPort and USB Type-A connectors for backwards compatibility (Figure 4 and Figure 5).
USB-C allows the use of an adapter or dock to offer DisplayPort- and HDMI-compatible video, analog and digital audio output and input, additional USB and Thunderbolt jacks, and Ethernet networking.
For convenience and to preserve multifunction USB-C ports, some computers retain DisplayPort ports (like some Dell computer models), while others may include one or more HDMI ports to allow additional displays beyond those supported over the USB-C bus, like Apple’s M1 Mac mini. Many still retain the standard RJ-45 Ethernet jack, too, as in the M1 iMac shown in Figure 4.
Most general-purpose computers have some or all of these ports:
Data: USB-C ports for input, output, and storage devices using the USB 3 or 4 or Thunderbolt 3 or 4 standards
Power: A power jack for an AC adapter; laptops may use USB-C
Video: HDMI or DisplayPort for connecting displays
Compatibility: USB Type-A ports for backwards compatibility, charging, or lower-performance needs, like keyboards and mice
Networking: Ethernet (1 Gbps or faster)
Camera cards: Camera card slot for SD or MicroSD
Audio: 3.5 mm stereo audio output
Consider Where Hubs and Docks Fit In
It wasn’t that far into personal computing that people realized they often needed more ports than a computer manufacturer provided. Computer makers considered the needs of the many, not the few; anyway, if you had an extensible machine, you could just buy cards and plug those in and go on your merry way.
Third-party companies and evolving standards filled this ecosystem niche, one that computer makers had less interest in. The answer to more ports might be a splitter, a daisy chain, a coordinating multi-port hub, or a dock. Hubs and docks remain in wide use today.
With early devices, standards weren’t always made to have one port accept the data from multiple devices. Clever folks created combinations of software drivers and hardware coordinators that allowed adding more and varied input devices.
Daisy chaining dates back long before personal computing. Each peripheral has one input and one passthrough port. The input port connects to the host device or previous peripheral in the chain; another peripheral plugs into the passthrough port. At varying times, daisy chaining was used with serial devices, printers, scanners, disk drives, MIDI instruments, and displays. Thunderbolt 3 and 4 still supports daisy chaining.
Hubs and docks provide a more fully supported and robust way to multiply ports. They don’t just fool a host adapter or relay signals, but they work with basic protocols to coordinate many separate devices using the same standard (usually called a hub) or split one output port into a number of different kinds of ports (usually called a dock).
Some computers in the 1980s to 2000s—mostly laptops—came with a specialized connector that was more like an external card slot, providing direct access to an internal bus. These docks would supplement what were often a handful of connections on a laptop, providing a full array of desktop-computing-style ports.
USB and Thunderbolt dominate modern hubs and docks. A USB 2.0 or USB 3 hub might be able to connect from four to a dozen devices all at high speed, and some ports might provide extra power for charging. A Thunderbolt dock could be festooned with ports—one currently on the market has eighteen. It’s only since 2021 that Thunderbolt hubs that add more Thunderbolt 3 or 4 ports began to appear.
Docks typically sport a variety of ports and are most valuable with computers or mobile devices that only include USB-C or a couple of ports on their own (Figure 6). Apple released laptop Macs from 2015 to 2020 with only USB-C ports, used for USB 3, Thunderbolt 3, video output, networking, and charging. (The company backed off in 2021 on being quite this parsimonious in port severity.)
With the rush to USB-C, many people find themselves short of USB Type-A ports and need either a hub or a dock. A dock can wind up making more sense, even though it will always cost more than a simpler hub, just for the range of what’s offered.
Docks may not just provide missing ports, but support standards not included on a computer that has many different kinds of connections. For instance, many computers lack HDMI and DisplayPort connectors, or only have one of those kinds; the older Mini DisplayPort, used primarily on Macs and a generation of Apple displays, disappeared long ago (see Figure 6). If you have a display with a hard-wired data cord or want to avoid adding a separate USB-C adapter, getting a dock with extra ports and the correct video port can be the right call.
Attach a Cable or Interpose an Adapter
Two ports will just stare emptily at each other across a void if they lack a cable to connect them. Similarly, a jack for one standard will look forlornly at an incompatible port it knows it should match with until an adapter is interposed that provides the translation service required.
Cables Glue Devices Together
A port is connected to a controller via a bus, but a cable is the glue that links a port on a host computer or mobile device to a matching peripheral, whether a standalone device, a hub, a dock, phone, tablet, or even another computer.
Cables may have the same plugs on each end or different ones. Until the 2010s—largely with the exception of phone-jack-style Ethernet—most ports/peripheral pairs used dissimilar ends: one for a peripheral or other device and one for a port on a host. It took the introduction of Thunderbolt with Mini DisplayPort and USB 3.1 via USB-C for that to change, seemingly forever. (I dig into that further in The USB Standard.)
In the current era, you are most likely to work with one of the following cables:
USB-C to USB-C: For USB 3.x or 4 data, Thunderbolt 3 or 4, power, or some combination
USB-C to USB Type-A: Host device to a peripheral with USB-C
USB Type-A to USB 3.0 Micro-B: Host to an external drive (as noted later in USB and Storage Jacks)
USB-C to HDMI or DisplayPort: Display connection
USB-C or Type-A to Lightning: For Apple mobile devices
Adapters Convert or Extract Standards
Many standards can be wrapped inside the data of another standard, something called encapsulation. Encapsulation lets docks work: a connection to a single port on a computer, say, can be split out by a controller on the dock to the physically different native ports for Ethernet, DisplayPort, HDMI, USB 2 or 3.x Type-A, and Thunderbolt.
I talked about docks above, but encapsulation also allows for adapters. An adapter is like a micro-dock: instead of multiple ports, it offers a single jack and plug, sometimes so small that it’s just big enough to contain a plug in its jack (Figure 7).
An adapter almost always includes a microchip (sometimes two) that performs one-way or two-way conversion. The same is true for a cable with dissimilar ends, like a USB-C to DisplayPort cable, which is an adapter in cable’s clothing.
You can find adapters that go much further than simple conversion, too. The most common example is the DisplayLink standard, built into adapters and docks, which allows video signals to be encapsulated over USB 3.0. DisplayLink’s creator, Synaptics, licenses this technology to third parties, who build it into all sorts of products. A DisplayLink software driver for macOS, Windows, and other platforms makes the adapter appear to be a standard video port which can pass DisplayPort or HDMI data to an external display.