The physical connectivity of the network is just as important as the data that flows across it. In this video, you’ll learn about the different devices that we use to build our physical network infrastructures.
<< Previous: Network TypesNext: Networking Tools >>
One of the devices that was first used to connect our networks together was a hub. A hub is often called a multi-port repeater, because there’s really no intelligence to how a hub works. Any signal going into one interface on a hub is going to be automatically transmitted to every other interface on this device.
A hub operates at half-duplex. That means of all the devices connecting to the hub, only one device can communicate at any particular time. As you might expect, that means that as the amount of traffic increases, the amount of throughput is going to be less and less efficient. You usually see a hub communicating at either 10 or 100 megabit, especially on an ethernet network. And because these are legacy devices and they’re not usually in our production networks, they’re more and more difficult to find today.
Instead of using hubs these days, we use switches to connect our networks together. Because switches are much more intelligent about how they send information across the network. These switches are bridges that are making forwarding decisions in a hardware called an Application-specific Integrated Circuit, or an ASIC. It makes its forwarding decisions based on the destination address. When you send data into a switch, it knows exactly where that information should go and sends it out a specific interface to that destination. This is very different than a hub, which sent the information to every single interface all at the same time.
These switches can be very small or they can be many hundreds of interfaces. And they might be in the core of an enterprise network. It’s also common to run power through these connections using Power over Ethernet if you have a switch that supports that. And these days, switches may even include additional hardware or software that allows them to do many things simultaneously. It’s very common these days to find a multilayer switch that cannot only provide switching functionality, but also includes router functionality.
If you are communicating from one IP subnet to another, then you will need a router. These routers make their forwarding decisions based on the IP address of the destination. There’s usually a routing table within the device. And the router is able to look up and find exactly where that information should be going.
If you have a switch that’s not only switching but also routing, it’s very common to hear this referred to as a layer 3 switch, referring to that OSI layer 3 functionality that’s commonly associated with routers. Unlike switches, routers can also connect many diverse types of networks. It’s very common to see local area networks, wide area networks, copper connections, fiber connections, and many others all connected simultaneously to a router.
Inside of our small offices or our homes, we probably have a wireless router. Inside of that wireless router is this Wireless Access Point functionality. This WAP, or Wireless Access Point, is effectively a bridge that connects everybody on this wireless network to the wired ethernet network. It makes its forwarding decisions very similar to a switch. It looks at the destination MAC address and sends information intelligently out to the wireless network or into the wired network.
You’ve probably heard the term modem before. But you may not know that modem is an abbreviation of modulator/demodulator. This modem is responsible for taking the analog sounds that you might hear over something like a phone line and converting them back to digital signals that a computer can understand. You would generally need a modem on both sides of the conversation to convert digital to analog and analog back to digital.
Many dial-up modems use the standard phone lines. Of course, there’s very limited frequencies over standard phone lines. And therefore, you get smaller amounts of bandwidths when you’re communicating over those phone lines. Those types of connections are called the Plain Old Telephone System, or POTS, modems. Those are the ones that we traditionally think about with our legacy computers. But these days we have newer modems, like ADSL or DSL modems, that we’re using to get higher speed communication over those phone lines.
Your network most certainly has a firewall connected to it, usually integrated into a wireless router or into a standalone firewall. These devices are responsible for filtering out traffic. And it’s usually making those decisions based on a TCP or UDP port number. Although, some of our newer firewalls can make forwarding decisions based on the applications that are running over the network and make its decisions on whether to allow or disallow that traffic based on the data inside the packets themselves.
Many firewalls, very similar to the one we have pictured here, has VPN functionality, the Virtual Private Network. So it can encrypt information that’s going into and out of your network, creating an encrypted tunnel between two sites.
Some firewalls can also act as a proxy. A proxy is a security device that makes the network communication for you. So you communicate to the proxy. The proxy makes the actual communication to the internet. And then once the proxy believes that that information is safe, it then provides you with that information. And in many cases, firewalls can also be routers. So instead of having a separate router and a separate firewall, it’s very common to combine these two devices together.
If you’re in a relatively large environment, you’re probably using patch panels. It’s very common to see patch panels that might have punch-down blocks or a combination of a punch-down on one side and an RJ-45 on another. Those patch panels making things much easier when you’re in the data center.
You usually have a single run from a person’s desk to the backside of a patch panel. You can almost see the wires in this picture that are on the backside of this particular patch panel. And they’re punched down permanently. That’s because those wires will rarely move once you have them in place. But on this side of the patch panel, where your switches and your routers and other devices are, you may need to disconnect and connect people to different devices.
So the patch panel becomes very easy to make quick connections between those devices that are out on the floor and the devices that are inside of your data center. In the case of this patch panel, in fact, you have a punch-down on one side. But on this side of the patch panel, it’s all RJ-45. So you can very easily disconnect and connect and change cables around, all without having to change any of the wiring between the data center and the end user.
One of the challenges you have with a number of technologies is that you often need them to go farther than what the standard specification allows. In these cases, you may want to have something like a line driver or an extender that’s able to extend the length that you would normally have, especially if it’s a copper cable. It’s very common to have an extension like this where you can plug in a copper connection and then extend it across to another one of these line drivers with a fiber connection.
It’s very common to see these line drivers and extensions used for things like ethernet connections or serial connections that have a limited length when you look at the copper connection. But by adding this fiber in between these line drivers, you’re able to take it to kilometers in length or more. These devices are powered. So you usually have a single power supply that you plug into the back. Or these might be rack mountable. They might have a single set of power supplies that covers many of them simultaneously.
Although these do provide us a way to extend the length of these copper connections, it also adds additional devices into the middle of a link. And of course, things can always go wrong when you’re adding more equipment. So you need to make sure once you get these in place that you test them very well and they’re able to send the information you’re expecting.
If you do a bad power supply, then the device isn’t going to work at all. And you’ll get no signal going over that link. And if you have a bad power supply, it may add additional interference to the connection. And you may have an intermittent connection or no connection at all.
A niche kind of network connection might be one like PLC, for Power Line Communication. You might also see this marketed as Ethernet over Power, or EOP. It’s an IEEE standard 1901. And the idea is that we’d be able to run standard ethernet connections over the power lines that are already in our buildings.
This standard supports speeds with ethernet of up to 500 megabits per second. And it’s very common to see this being used for things like the final link into a premise. Maybe you have it within a single building. Perhaps you’re connecting smart devices that you have to plug in and charge. Why not also connect them to the ethernet network at the same time?
Power over Ethernet is the exact opposite of PLC. Instead of putting ethernet on a power line, we’re instead putting power on an ethernet connection. This is very common to use when you have devices that need power but may not necessarily be anywhere near a power connection. Or maybe it’s for convenience. You may want to power your telephones or power a camera without running a separate power connection to that device.
You might have an ethernet switch that already has Power over Ethernet built into the capabilities of the switch. But if your switch doesn’t provide Power over Ethernet, then you may want to put an injector between the switch and the end device. Here’s a picture of a Power over Ethernet injector that I’m using to power a camera. If it’s powered at the switch, we call that an endspan. If it’s powered in the middle of the connection, then it is a midspan.
We generally see two different types of power modes for Power over Ethernet, or POE. Mode A means that we are going to put power on the exact same wires that we’re using for data. With mode B, we’re using some spare wires inside of the cable that aren’t being used for data. And we’re going to use those exclusively for the power. You’ll need to look at your end devices and see what type of power modes they’re expecting to be used when you’re taking advantage of Power over Ethernet.