Routers

Because each network card on a network has a different MAC address, each network group is assigned a network number. This network number, which defines the “area” of the network, is similar to a telephone area code. Different network area numbers are connected via a router, which, as you might expect from its name, routes packets from one network area (routing domain) to another.

---

Router is pronounced row-ter, not root-ter.

---

You can think of a router as the mystical piece of equipment in the telephone network that connects you from, say, New York to Georgia; routers usually connect geographically separate networks (this isn’t always true, but it’s a reasonable rule of thumb). More specifically, a router is connected to more than one network number, and it helps network calls get where they need to go (assuming the language in the network call is supported by the router).

---

This is where the analogy breaks down: Although phone switches care nothing about the conversation on the wire, routers do care. It might help to think of routers as friendly neighbors who can pass your message along (that is, if they speak your language).

---

Of course, the phone system needs equipment peppered throughout your neighborhood, and it also needs the high-end equipment at the telephone company. The same goes for your network. Two types of network “glue” tie you and your network neighborhood together:

•  Hubs (concentrators)

•  Switches

Think of a hub as a shared neighborhood party line (like the one you used to have at Uncle Harry’s cabin in the woods). It used to be cost-prohibitive to give each network card its own channel to talk on, because more silicon “smarts” were required. Just like its name implies, a hub is where all of the NICs on the “party line” (or segment) come together. Each NIC’s wire is connected to the hub in a spoke-like fashion, where they all physically share the same wire.

---

A common physical connection like this is generally referred to as a bus, and each participant that tries to use this common connection “gets on” or “gets off” the bus.

---

Because NICs share a common bus, each NIC can “hear” other NICs. NICs have built-in rules for using this common area, which is discussed further in Hour 9, “Ethernet Basics,” and Hour 10, “Token-Ring Basics.” For the moment, you just need to know that each NIC has been schooled in “netiquette” and has been taught in the factory to play well with others and to share the wire nicely—that is, most of the time. Be aware that if a network card does not obey these rules, this can cause problems on the entire shared segment. (See Figure 1.2 for an illustration of shared versus dedicated telephone lines and network segments.)

Figure 1.2  Shared versus dedicated lines.

---

Computers on older networks were actually connected together on the same wire, much like older Christmas lights used to be. Therefore, a physical break in the wire meant that all the computers on one side of the break went down. A hub-based segment fixes this, because each computer has its own physical connection, and the hub ties them all together. The newer method of sharing the wire is called a star topology, because it looks like a star. The older method is called bus topology—the line from PC to PC is the common bus area.

---

A new technology called switching allows each network card to have its own private line; however, a lot of hub-based shared networks are still in existence (and don’t look for them to go away any time soon). You’ll find switching to be more and more exciting as you get into network troubleshooting. Similar to the way hubs allow users to get away from the “Christmas light bulb” problem, switches allow users to get away from many of the problems of shared networks. Instead of each computer having to compete for the right to talk when it wants to, each computer can transmit pretty much when it needs to—as long as the computer on the other end is available, of course. Others in the network neighborhood no longer compete for the right to talk on the wire, because each computer has its own connection.

How does this work? In a nutshell, a switch looks at the MAC addresses of two workstations on the network that want to communicate, and it opens up a high-speed private channel for them. A hub has one common channel, but a switch has many, many channels, as well as the intelligence to switch conversations between them. With a fast enough switch (called a wire speed switch), workstations conversing on the switch are not affected whatsoever by other workstations talking away at full speed.

Although prices on hubs are very low, switching has become very cheap as well; you can look to this technology as the future of local area networks. Although shared hubs have their virtues (for example, it’s easy to see the conversations on a shared line, which is good for troubleshooting), the fact remains that they are more trouble-prone than switched networks (switched networks are much faster and more reliable). As switches get even cheaper, you can expect to deploy them instead of hubs.