Hi everybody.
This discussion is based only on Cisco inline method.
Suppose we have a cisco ip phone connected to a switch. Switch is usig cisco inline method to deliver power over ethernet.
ip phone----------------------f1/1SW
Everything is working smoothly and ip phone is getting its power until a user decides to replace the ip phone by some other non PoE device.
How will cisco inline method combat such situation?
I tried to find answer to this design issue and came across this excellent white paper.
http://www.cisco.com/warp/public/cc/so/neso/bbssp/poeie_wp.pdf
However, I am left with more questions. I quote from the above link.
Periodically, the PSE will check to see whether the powered device is still present and requires power and also
implements checks to detect conditions such as where a short circuit may have been placed between transmit and
receive pairs. Cisco implements two mechanisms to detect these conditions. The first is an extension of the
prestandard discovery protocol whereby a discovery signal is transmitted periodically. If the received discovery signal
has the same amplitude as the transmitted signal, the PSE will remove power, as there is a short circuit. If the PSE
receives a discovery signal that has been attenuated by the low pass filter, the PSE will maintain power to the powered
device.
So Cisco inline method periodically sends discovery signals. According to above paragraph, it will help detect two problems:
If there is short.
If PoE device is still connected.
Pretty simple until I came across the following in the same link.
To prevent users from swapping patch lead connections between powered and nonpowered devices, a mechanism is
required to detect that a powered device has been disconnected and power delivery withdrawn from the port before
a nonpowered device can be attached. The IEEE 802.3af-2003 standard defines two mechanisms for disconnecting
power once a device has failed. Both mechanisms provide a solution to protect against a user unplugging a powered
device and plugging in a nonpowered device that would potentially damage the nonpower device. It should be
remembered that although two mechanisms are described, they provide the same functionality albeit in a slightly
different way. Cisco prestandard Power over Ethernet ports have a power disconnect mechanism that will remove
power from the port if the Ethernet link status is down.
So Cisco inline method removes the power from the port if the ethernet link status is down.
Let revisit my scenario.
If a user replace cisco ip phone by non PoE device, the ethernet link status will go down which in turn causes the switch to remove power from the port. Pretty simple !
Compare this to first excerpt where it says switch sends discovery signals to ensure PoE device is still connected. Why does switch have to send discovery signals periodically because if some one does replace PoE device by non PoE device, the ethernet link status will go down which causes switch to remove power? My reasoning is since switch can determine if PoE still connected or replaced by some non PoE by checking ethernet status, It therefore eliminates the need on the part of switch to periodically send discovery signals.
I understand the need of periodic discovery signals on all inactive ports .But sending discovery signals periodically on active ports does not make any sense to me. The key word is" Periodically". It makes completes sense to me to send discovery signal on a port when the port first become activated to see if the connected device is PoE or not. But once it is determined that connected device is non PoE, there should not be any need to send periodic discovery signals because if someone does swap non PoE device with PoE device, the ethernet link status will go down and go up which will cause the switch to send discovery signal and thereby detecting PoE device.
Sorry for this long winded post.
I will appreciate your input.
thanks.
