Networking

Get IT Done: New uses for 802.3af devices on your network

Power-over-Ethernet gains new life with todays technology

Have you ever noticed that during a blackout, you can pick up the telephone and call the power company to report it? This is possible because the telephone line carries enough electricity to power a landline telephone (not, of course, a cordless phone or phones with built-in answering machines). Obviously, the idea of providing power over a communications line is a proven technology. After all, the telephone has been around in one form or another for over 100 years. The basic principle of providing power over a communications line has also been applied to some Ethernet devices, using a technology called 802.3af.

What’s 802.3af?
Believe it or not, 802.3af technology has been around for a while; it just hasn’t really caught on yet. The first implementation that I ever saw of 802.3af was through something that 3Com called Power Base T. Back around 1999, I bought my first wireless access point, a 3Com Air Connect. Because the unit had limited range, the instructions recommended mounting the unit in one of the building’s corners near the ceiling. Naturally there are few electrical outlets near most ceilings. This is where the Power Base T module came into play.

To power the unit, it was necessary to run a standard Cat 5 cable to the Power Base T module and connect the wireless access point’s power supply to the Power Base T unit. From there, you would run a second Cat 5 cable from the Power Base T module to the wireless access point. The electricity needed for powering the access point traveled across an unused pair of wires in the Cat 5 cable. The result was that you could place the access point in any location accessible via Cat 5 cable without having to worry about whether an electrical outlet existed at that location.

Using 802.3af today
OK, so the year isn’t 1999 anymore and most of the wireless access points out there don’t require you to place them in obscure and difficult-to-reach locations. So why would anyone use 802.3af? There are actually several different ways businesses can benefit from rolling out this technology.

First, there’s the simple reason that you’d want to reduce the number of devices that you have to plug into a wall outlet. The average employee usually has a PC on his or her desk. Naturally, the PC and the monitor must both be plugged into an electrical outlet. Additionally, an employee may have a scanner, printer, or other peripheral attached to the computer. Once again, this means more devices that must be plugged in. Add a docking cradle for a PDA, a charger for a cell phone, and a cool IP phone, and you’re really starting to plug in a lot of devices.

Eventually, an employee may find that he or she has more devices than a traditional surge protector can comfortably accommodate. When that happens, the company will have to either daisy chain surge protectors as a way of giving the employee more electrical outlets, or they will have to install more outlets.

But daisy chaining surge protectors together can cause problems. If employees do too much daisy chaining, they may find themselves tripping a lot of circuit breakers. Using 802.3af devices eliminates at least one more thing that must be plugged into a power strip, potentially overloading circuits.

Another advantage is similar to the scenario mentioned above: 802.3af allows you to locate devices where you don’t have electrical outlets to begin with. Hopefully, when your offices were designed, there were plenty of outlets planned. However, if you later discover that you need an outlet, then you’ll have to pay an electrician to install one. With an 802.3af device, all you need to do is plug the device into the nearest network jack.

What can I use 802.3af for?
Let me start by saying that you can’t expect to power a desktop computer off of an 802.3af power supply. However, lower voltage network devices are fair game. Some examples of devices that can be powered by 802.3af connections include laptop computers, Web cameras, wireless access points, and IP phones.

As a general rule, anything that requires a network connection and that doesn’t use a whole lot of electricity could be theoretically made to work with 802.3af technology. The biggest thing to remember is that before you can plug an 802.3af connection into a device, the device must be specifically designed to accept 802.3af power.

If you were to connect 802.3af power to a non-802.3af compliant device, you would most likely cook the device’s NIC, if not the entire device. Some of my Internet-based research sources indicate that it’s safe to connect an 802.3af module to a non-802.3af compliant device, but personally I would not take the chance with an older 802.3af device.

Newer 802.3af devices eliminate this problem completely in a rather fascinating way. Each of the newer 802.3af-compliant devices contains a digital signature that designates it as an 802.3af compliant device. Whenever a device is connected to an 802.3af hub or switch, the hub or switch looks for the signature on the device. If no signature is detected, then no power is sent to the device and the hub or switch services the connection with a standard, nonpowered connection, just like the ones used by any other hub or switch. If, on the other hand, the digital signature is detected, then the hub or switch sends power to the device as is required.

Powering multiple devices
Today’s 802.3af units support point-to-multipoint connections. This means that the 802.3af power supply can be integrated into a hub. When you connect a network device to the hub, you aren’t just connecting the device to the network, you are also powering the device by transferring some of the hub’s power to the device over the Ethernet connection. This means that you can power multiple devices through a single hub. Since the hub is a single device, it only requires a single electrical outlet. Therefore, rather than consuming dozens of electrical outlets for these small network devices, you’re only using a single outlet for the hub.

The point-to-multipoint architecture used by 802.3af hubs has another benefit too. As you probably know, most businesses connect their hubs to a UPS so that if the power goes out, network communications can continue to function. The problem is that only PCs that are also connected to a UPS will be able to take advantage of the still-functioning hub. Because a decent UPS tends to be so expensive, most of the companies that I’ve worked with tend to place servers on UPSs, but don’t connect the PCs to the UPS.

If you’re using 802.3af though, this works a little differently. If your 802.3af hub is connected to a UPS, then the devices being powered by the hub are also technically connected to that same UPS. Therefore, a single UPS could provide power to dozens of devices through a single hub.

The 802.3af standard defines two different types of 802.3af implementations: end span and middle span. End span is the type of device that I was just talking about in which the power is passed to the device through a hub or switch. Middle span is like the Power Base T module discussed previously. In this implementation, the Power Base T module sits between the hub and the network device.

Specifics of the 802.3af Architecture
So far I have explained what the 802.3af standard is and what it can be used for, but I haven’t given you a lot of specific information on how much electricity can really be passed through an 802.3af connection. As you may recall, earlier I mentioned that you could power small devices, such as IP phones or wireless access points, through an 802.3af connection, but that you wouldn’t be able to use that connection to power something big like a desktop computer.

Part of the reason for this is that 802.3af passes the electricity to the device over a standard Cat 5 Untwisted Pair (UTP) Ethernet cable. This cable was designed to carry data, not electricity, so you are very limited in how much electricity can be safely passed through the cable.

As you may know, a Cat 5 cable contains four pairs of wires (eight wires total). A standard Ethernet device only uses two of the four pairs (four of the eight wires) and thus leaves two pair free for other things. 802.3af uses the remaining two pair (four wires) to carry electricity. The amount of electricity used must be low enough that it won’t burn through the thin wires, and must be small enough that it won’t disrupt the data signals carried by adjacent, unshielded wires.

The wires used depend on the 802.3af implementation. If an end span is being used, then transmission pairs 1-2 and 3-6 are used to carry power to the device. If a mid-span implementation is being used, then power is sent along pairs 4-5 and 7-8.

As far as the amount of power sent to the devices, there is a lot of contradictory information on the Internet. According to some sources, a standard amount of power is used for each connected device. Other sources indicate that the standard amount of power is actually the maximum power level and that some devices can actually tell the hub or switch how much power (at or below the maximum level) to use.

In either case, the maximum amount of power that can be delivered to an 802.3af device is 350 milliamps at up to 12.95 watts. This takes into account the power loss caused by sending the power across long cable spans. Just to give you an idea, a wireless access point generally consumes between 3.5 and 10 watts of power depending on its design and range. If you would like to know more about the 802.3af standard, check out this 802.3af requirements Web site.

One other feature
There is at least one more cool feature found in 802.3af devices. Earlier, I mentioned that some of the newer hubs and switches can check a connected device for a digital signature and can then send the appropriate amount of power to the device once the signature has been authenticated. You might have wondered how this works being that the device is unpowered at the time that the signature is checked. As it turns out, there is actually quite a bit of intelligence built into newer 802.3af-compliant devices.

You may have heard of a technology called Wake on LAN. Wake on LAN is used to force a PC to power up when a request is sent to the PC’s NIC. The electronics that allow an 802.3af hub or switch to check for a digital signature are very similar to those used in Wake on LAN implementations. The biggest difference is that while Wake on LAN uses some of the phantom power that is always flowing through a system board that’s powered by an ATX power supply, the 802.3af device has no power. Therefore, power detection is done by sending a “safe” amount of power through the normal power supply lines. As soon as the detection process completes, power is either cut completely or applied fully.
0 comments

Editor's Picks