IP Telephony Troubleshooting

Enterprise Networks & Servers, Dec 2004 by Trunk, Ronald

Now that IP telephony systems are becoming commonplace, network engineers need to be familiar with IP telephony operations and troubleshooting procedures. IP telephony is almost always a mission-critical application, so keeping the system running smoothly is an important task.

A thorough understanding of the interaction between the various components of the IP telephony system will help you quickly identify and resolve most common problems. Your users expect the IP telephony system to be as reliable as a traditional PBX, so resolving problems quickly is important if your users are going to accept the system.

If you're responsible for maintaining the network infrastructure (routers, switches, etc.) you need to be able to tell whether the problem is network-related, or whether it is a problem with the PBX configuration. Often, the staff responsible for the network infrastructure is not the same that is responsible for operating and maintaining the PBX functions (adding users, maintaining voicemail, etc.). So it's important to get the right resources working to solve the problem (Note: that's not the same as finding someone to blame!).

In a traditional PBX, each telephone is directly connected to a physical port on the PBX. The phone communicates directly with the PBX for all the signaling, audio and power. This makes troubleshooting fairly easy; either the phone is communicating with the PBX or it's not. If it is, the problem is probably a PBX configuration problem. If it isn't, the problem is often simply a physical connection problem between the phone and PBX.

In IP telephony, things get a little more complicated. There are a number of components that all have to work together correctly. The connections between them are logical, not physical - and transitory at that - so troubleshooting problems becomes a lot more involved. You need to understand the interactions between the components in order to find the problem quickly.

The components of an IP phone system can be divided into two groups: The phone components themselves and the network infrastructure components. The phone components are the IP phones themselves, the IP PBX, various gateways that provide the interface between the IP network and the telephone network (the PSTN), and other devices such as conference bridges, gatekeepers, voice mail servers, etc. The IP PBX is a server running a PBX application such as Cisco's Call Manager, Avaya's Communications Manager, or Nortel's Succession. There's even an open source product called Asterisk.

The network components include the infrastructure itself (the routers, switches, wiring, etc.) and also ancillary services that are important to IP telephony: DHCP servers and TFTP servers (Note: if you have more than a handful of PCs and are still using static IP addresses instead of DHCP, we need to talk!).

Understanding the Boot Process

To begin, you should be familiar with the boot process of the IP telephone. A failure to boot properly is often a result of some network-related problem. Understanding how a phone initializes will help you see what can go wrong and how to fix it.

We will look at the three major vendors of IP phones: Cisco, Avaya and Nortel. At a high level, all three follow the same basic steps.

1. The phone requests an IP address from a DHCP server.

2. The DHCP server gives the phone an IP address and additional information, such as the address of the TFTP server and/or IP PBX.

3. The phone connects to the TFTP server and receives configuration information and/or updated firmware.

4. The phone connects to the IP PBX and registers itself with it.

5. The IP PBX configures the phone's operational features, namely the phone number, programmable buttons and more.

When a phone has successfully registered with the IP PBX, the phone's display will show the extension number and usually the time of day. If you see this on the phone you know a couple of things.

1. The phone received a valid IP address from the DHCP server.

2. The phone was able to exchange data with the TFTP server.

3. The phone was able to exchange data with IP PBX and the PBX was configured to recognize the phone.

If a phone fails any of these steps you will not see the "normal" display. Instead, you will see various status or error messages indicating that one of the above things did not happen.

To understand what went wrong, let's look at each of the vendor's phone systems in a little more detail.

Cisco

When a Cisco phone is powered up, you will see the mute, headset and speakerphone indicators light up. Next, the phone and switch exchange Cisco Discovery Protocol (CDP) data. CDP is a Cisco proprietary protocol that lets interconnected Cisco devices learn about each other. The switch uses CDP to tell the phone which VLAN will be used for voice traffic. (Note: if you are not using separate VLANs for voice and data traffic, we need to talk!) During this process the phone will display "Configuring VLAN."

Once the phone knows which VLAN to use, it will tag all of its data frames with the appropriate 802.1q tag and request an address from the DHCP server. The phone will display "Configuring IP." The DHCP server will offer an IP address as well as the address of the TFTP server.