What cable modems, voice over IP terminals, and any other devices that communicate with a headend device known as a “CMTS” (Cable Modem Termination System) do when they’re activated for service: Seek connection with their “parent” (the headend device), to be recognized and receive instructions. Similar to what a cell phone does every time you turn it on. In the case of the cable modem, the process of ranging is preceded by a “registration” ping — the equipment equivalent of “are you my mother?” — sent in the 5-42 MHz upstream path of the cable plant, to that headend CMTS.

Once the CMTS recognizes the cable modem, ranging begins. This is where the CMTS and cable modem establish how far apart they are, in distance, as a function of time. The two exchange pings, and from them extract timing information, to learn distance and to synchronize with one another. That way, the CMTS can be more intelligent when optimizing bandwidth usage, for that modem and any others connected.

The software, usually within a billing system, that establishes what rate to apply to a transaction, based on pre-established business rules. For instance, a telephony ratings engine calculates the price of a call, based on time of day and rate per minute. Or, in a multiple ISP environment, the ratings engine reinforces contractual agreements about bandwidth usage. In a spreadsheet sense, the work of the mediation engine determines what’s in the cell; the ratings engine is the equation applied to that cell, to come up with the answer.

The process of removing and replacing the physical components of telecommunications plant, such as strand wire, fiber and coaxial cable, amplifiers, nodes, passives and connectors, so as to add capaicty. The word itself is used as both a noun and a verb. As a noun: “Our 750 MHz rebuild…” As a verb: “As soon as the ground thaws, we need to rebuild …”

The rebuild is one of three types of activities that physically affect the broadband infrastructure. Another is the “upgrade,” which involves replacing some of plant components, like amplifier modules, but doesn’t usually require ripping out existing plant. The third is “new build,” also known as “line extension,” which is what happens when you need to drop a line into a new neighborhood.

From a capital expense perspective, rebuild falls in the middle. New builds are most expensive; upgrades are least expensive, on a cost-per-mile basis.

A code for correcting transmission errors in digital communications signals, and particularly delivered over cable digital networks. Reed-Solomon encoding works by assigning 20 “code” bits for every 187 information-bearing bits. The coded bits are correlated with the information-carrying bits, so that if a noise spike obliterates a chunk of traveling data, the Reed-Solomon algorithm has a basis for rebuilding the deconstructed stream. Named for Massachusetts Institute of Technology scientists Irving Reed and Gustave Solomon.

One of several terms to describe an encryption and conditional access system that can be physically removed from a device. The most well-known example of removable security is the CableCARD, which slides into a slot in a digital or HDTV to descramble premium cable services. Removable security became considerably more mainstream in 2007, which is when an FCC rule (thrice delayed) went into effect, stipulating that cable operators provide removable security in all going-forward digital boxes.

A catch-all term, loosely synonymous with “home media center,” “network interface unit,” or “home network.” In general, then, the unit that joins the broadband network (cable or DSL) with various consumer electronics devices inside the home. To know the “residential gateway” is to first determine who’s talking about it, and which industrial segment they represent. In general, the residential gateway describes an electronics device, developed for home use, to deliver broadband packages to consumers. As of early 2001, at least six distinct industries were vying to “own” the latch to this broadband gate: On the service provider side, 1) telcos, with DSL and IPTV techniques; 2) cable, with broadband; and wireless. On the equipment side, 1) Telecom, 2) Data/computer and 3) cable industry suppliers.

The portion of a bi-directional passageway that carries information from a subscriber’s home, back to a signal collection and processing center. In cable broadband systems, the return path is synonymous with “reverse path” and “upstream path.” The signal collection center is the headend; “bi-directional” is synonymous with “two-way.”

The U.S. cable return path spectrum is located between 5 MHz and 42 MHz — although often it stops as low as 30 MHz — depending on equipment and plant conditions. Spectrally speaking, the 37 MHz or so of bandwidth that embodies the return path is an inhospitable zone, highly susceptible to two types of noise: Ingress (signals leaking in) and impulse (electrical spikes.)

What makes it even worse is this crazy little fact: Most noise — upwards of 70%, by some estimates — originates in the home. That noise mixes with the intended signal as it moves upstream, and gets amplified along with the intended signal as it makes its way to the headend. So the desired signal and noise from eight homes goes to the tap, which gets joined to other pockets of eight homes, before it gets to the amplifier; at the point of the optical node, the noise and intended signals from as many homes are connected to the node are coming in. This effect is called “noise funneling.”

The harsh conditions in the return path usually require a sturdy modulation type, such as QPSK (quadrature phase shift key), which is slower than downstream modulation formats (like QAM), but withstands noise better. This is not unlike driving on a road pocked with deep potholes: You have to go slower, so as not to damage your car (or your head, on the ceiling). On a smooth road, without potholes, you can drive faster without apparent road risk to your vehicle. Oversimplified, the return path is the bumpy road; the downstream path is the smoother road.

Nonetheless, the return path is critical to two-way applications, which require a way for subscribers to interact, instead of just passively receiving TV pictures or broadband services.

For the first few decades of cable television, the return path wasn’t needed. Television signals were broadcast downstream, through the plant, to homes. Subscribers turned on their TV, and watched. They didn’t “click” to initiate or pause or rewind a TV show, or to invoke a bound application. They didn’t use PCs to connect to the Internet — and so on.

In the late ’70s, some operators experimented with television and data services that encouraged consumers to interact. At that point, attention started to focus on building a two-way path to augment the existing one-way, downstream plant. Mostly, it involved the installation of modules into existing amplifiers that fed a signal upstream, to the headend, then balancing the two-way signal path.

These days (2005), cable’s return path is viewed as a critical competitive differentiator against satellite TV providers, because it’s kind of hard to create a return path from millions of homes, to a satellite, to destination. It’s what’s propelling VOD to be a time shifting mechanism for more and more types of television, rather than being contained to “digital PPV” only. It’s what gives passage to just about any form of interactivity, save what can be done with a built-in DVR. In simple, competitive terms, cables return path is what will likely go farthest to attract and retain customers, at least relative to satellite. (Meaning, telco video also comes with a return path, so that fight will be fought on different terms.)

A swath of electromagnetic spectrum that sits between 60 kiloHertz and 300 GigaHertz, and is used for telecommunications, TV and radio transmission. (Note that engineers from different industry disciplines tend cite different upper and lower bounds, depending on their specific focus.)

Generally speaking, the electromagnetic spectrum is very, very wide. Radio frequency is one chunk of it. RF begins just above the range at which your ears can hear, and ends at light. Above it is infra-red, visible light, ultraviolet light, soft and hard x-rays, and gamma rays, for example. So far as we know, the electro-magnetic spectrum is not infinite, but this remains a chewy debate among physicists.

Its called “radio” frequency because its introduction into engineering lexicon pre-dates television or broadband, back to the early days of ship-to-shore radio communications.

The hardware and software that makes a physical junction between two or more networks: A corporate network and a managed backbone, an Internet Service Provider (ISP) to the Internet, or a corporate network to a sister network.

Routers are the post offices of today’s digital, packetized world. If each packet is a “letter,” the router is the place where that letter is examined for destination and appropriate postage, sorted, and sent along its way. Just like a letter may move through several post offices before it gets to its intended recipient, a packet can move through several routers before it arrives at its intended destination.

Routers are built to be smart and to keep learning. Over time, for example, they learn where neighboring routers are, and how long it takes data to get to and from them. From learning, routers can assure data takes the best path, and can induce alternate routes if it has learned that a neighboring router usually gets pretty clogged at a certain time of day.

In cable modem parlance, the “Cable Modem Termination System,” or “CMTS,” is essentially bridge between two disparate networks — the HFC plant, and the Internet — linked to a router. Its “ports” are shared among individual cable modems on the one side, and connected to a high-speed Internet link on the other side.

A set of software commands written tightly to instantaneously process instructions, usually to manage the internal resources of an electronics device. Real time operating systems, such as those used in digital set-top boxes, reside “at the metal,” and are the dispatchers of orders to chips designed to execute those orders. Pronounced as a word: “are-toss.”

Real time operating systems are important whenever latency is undesirable. When you’re changing a channel on television, do you want a set-top to just change the channel — or do you want to see the hourglass while it “loads” the channel, like what Windows does when you load Word?

To run in real time, operating systems need to be unencumbered. Brevity matters. Less is more. They don’t want to be bothered checking in on other software programs that may also be resident. Some critical portions of the device’s existence are written into a “kernel” — a module — then burned into ROM (read-only memory) to be executed each time the device powers on. Other task lists are often streamlined into modules, too, to increase speed.

In the lasagna of set-top software layers, the RTOS is usually at the very bottom. Part of its purpose is to isolate the stuff below it — memory or processor chips, tuners, infrared transceivers, modems — from the layers above it. Middleware sits above the RTOS. Any unbound interactive applications are up top.

Note: The RTOS itself also consists of layers. The stacking varies, predictably, from supplier to suppler, but in general, bottom-to-top goes like this: Device drivers/Abstraction Layer, Core.

Generally, then, the undermost point of the RTOS is the “abstraction layer,” which collects, organizes, and retains information about individual hardware components. How? Chips report about themselves through permanently-stored lists called “device drivers.”

Scientific-Atlanta’s “PowerCore,” Sun Microsystems “JavaOS,” Wind River’s “VXWorks,” ISI’s “pSOS,” Microware’s “David,” Microsoft’s “Windows CE” (where the “CE” stands for “Consumer Electronics”), and Sony’s “Aperios” are all examples (living and dead), of real-time operating systems developed or adapted for use in advanced digital set-top boxes.