Understanding Voice Clients and Voice Traffic
Voice, like all other information, travels in packets over IP networks with fixed maximum capacity. The major difference between voice and data traffic is the fact that data packets can be re-sent if they are dropped, and then applied to the empty spots in data, thereby producing complete information. With voice, there is no point in resending packets because voice only makes sense in a stream of contiguous packets. Because resending voice packets is not an option with voice over IP (VoIP), voice traffic must be more carefully configured.
This topic applies to the transmission of voice, multimedia sessions, and video, which are all part of VoIP.
This topic describes:
What Is Voice Over IP?
VoIP refers to the communication protocols, technologies, methodologies, and transmission techniques involved in the delivery of voice communications and multimedia sessions over Internet Protocol (IP) networks. VoIP systems use session control protocols to control the setup and tear-down of calls as well as audio codecs that encode speech, allowing transmission over an IP network as digital audio using an audio stream.
Voice over Wireless Phones
A VoWiFi phone is similar to a cell phone except that the radio is an 802.11 radio instead of a cellular radio. VoWiFi phones are 802.11 client stations that communicate through an access point.
VoIP is also available on smart phones and Internet devices, making it possible for users of portable devices that are not phones to place calls or send text messages over 3G, 4G, or Wi-Fi.
Private Branch Exchange (PBX)
Private Branch Exchanges (PBXs) are telephone exchanges that serve a particular business or office. The PBX makes the connections for internal phone calls and provides a way to dial out to the public network.
How Is Voice Traffic Different From Data Traffic?
Voice transmission has different issues than data transmission because voice traffic cannot be re-sent if it is lost. Three unique problems that affect call quality are latency, jitter, and packet loss.
In addition, voice traffic can have roaming issues because people are more apt to walk around with phones.
Latency, Jitter, and Packet Loss Affect Voice Traffic
Latency, or delay, is the amount of time it takes the sound of your voice to reach the ear of the other person. Maximum acceptable latency limits for VoIP are 150-200 ms, depending on your call quality requirements. Delay levels that exceed 80 ms are indication of delay issues.
Jitter is the variation in delay between packets. Because packet jitter always varies, VoIP phones use jitter buffers to smooth out the variations. A jitter buffer is simply a First-In, First Out (FIFO) memory cache that collects the packets as they arrive, forwarding them evenly spaced and in proper sequence for smooth playback. Increasing jitter buffer size can help with jitter, but only to a point. Significant jitter might cause the jitter buffer to increase to the point that delay reaches unacceptable levels. Jitter levels that exceed 20 ms are indication of jitter issues
Packet loss occurs when the maximum delay specified in the jitter buffer is exceeded. Networks tend to either occasionally drop single packets (called gaps in packet loss), or large numbers of contiguous packets in a burst. Packet loss above 5% is considered unacceptable when using the G.711 codec—sustained bursts of packet loss cause the most problems.
Voice Traffic Is Susceptible to Roaming Issues
A key advantage of voice over WLAN is mobility, but voice roaming is potentially incompatible with 802.11i. In accordance with 802.11i, users traversing a network negotiate new encryption keys with every access point they encounter. If this does not happen quickly enough, voice quality is affected.
The 802.11i fast roaming specification eliminates delays associated with re-authenticating roaming clients. Instead of generating a new encryption key with each access point, the VoIP client can use the same key because the pairwise primary keys (PMKs) used are cached at the controller.
What Protocols Are Used for Voice?
Voice solutions require two types of protocol. Signaling protocols are used during call setup, management and teardown. These protocols generally require low bandwidth, might use a connection-oriented model (TCP), and are typically not delay sensitive. Bearer protocols actually carry the stream of voice samples. Bearer protocols are delay sensitive, are connectionless (UDP) and require special treatment to ensure prioritization over other types of traffic. A separate stream is usually required in each direction. Spectralink Radio Priority (SRP) is a legacy bearer protocol that you can configure.
What Is Different About Configuring for Voice Traffic?
Voice traffic needs more bandwidth than data traffic, and that bandwidth needs to be protected so it remains constant and prevents degradation. For these reasons, voice traffic works best when the WLAN Service profile (SSID) and Radio profile used are dedicated to and designed for voice traffic. We recommend that you separate voice and data by SSID and preferably by RF band. Also, be sure to ensure adequate coverage at -60 dBm to -70 dBm level and capacity for the expected voice load.
To optimize voice, also follow these suggestions:
Consider Using Call Admission Control (CAC) to Limit Clients per Access Point
Call admission control, configured in the WLAN Service profile, limits the number of concurrent phone calls. Some VoIP devices use Wi-Fi Multimedia (WMM) to provide call admission control for voice clients and some devices use Spectralink voice protocol SVP to provide call admission control for voice clients. Both methods work—which one you use is dictated by your voice clients. The only difference between WMM and SVP is that SVP gives highest priority to Spectralink traffic and WMM does not. If your network has both WMM (most common) and Spectralink voice devices, we suggest that you provide a dedicated SSID for each type.
Ensure That Network Equipment Supports Seamless Roaming
Voice traffic must use seamless roaming so no call is interrupted. You can achieve seamless roaming with:
At least 20% overlap between access points
Between -60 dBm and -70 dBm signal strength wherever voice is required
Roaming achieved within 50 ms
Minimized number of router hops between the handsets and the PBX
Support Quality of Service on all Hardware Used for Voice
QoS should be supported end-to-end with voice traffic. Ensure that infrastructure switches and routers support and preserve data prioritization across the full path of any voice streams.
Use Automatic Power Save Delivery to Preserve the Battery Life of Phones
Handset battery life is a major problem. Unscheduled–using Automatic Power Save Delivery (U-APSD) provides dramatic improvements in battery life, for example, from 2 hours talk time to over 10 hours.
Handsets that support WMM/802.11e are increasingly available.
Create a Unique WLAN Service Profile for Voice
Create a unique WLAN profile for voice—see Creating and Managing a WLAN Service Profile.
Create a Unique Radio Profile for Voice
Create a unique Radio profile for voice—see Creating and Managing a Radio Profile.