What to look for in voice over frame relay CPE
 
|
|||
 | |||
|
Sign up to receive this and other networking newsletters in your inbox.
In a previous issue, we attempted to dispel some of the myths surrounding voice over frame relay (VoFR). Today, we'll detail the capabilities of VoFR. Keep in mind that there are other voice over data technology options, such as voice over IP and voice over ATM, that can do many of the same functions as VoFR. When considering the VoFR equipment on the market, look for vendors who offer the following capabilities:
Compression: Unless you have access to cheap bandwidth, your VoFR equipment should compress your voice signal from 64K bit/sec to at least 32K bit/sec. In most cases, compression to 16K or even 8K bit/sec is possible and preferred. Some equipment vendors support dynamic compression options. When bandwidth is available, a higher voice quality is achieved using 32K bit/sec, but as other calls are placed or other traffic requires bandwidth, an 8K or 16K bit/sec compression algorithm is implemented.
Fragmentation: VoFR equipment must segment data and voice packets into small ones. This allows the voice quality to remain high. If fragmentation is not done, a large data packet may "block" or delay voice packets behind it. By fragmenting the packets, the voice and data packets will have a more constant delay and a reduced possibility of being temporarily blocked by a large data packet.
Prioritization: Real-time voice traffic is very sensitive to delay, while many data applications such e-mail are more delay-tolerant. Your vendor should support voice prioritization. If you are running delay-sensitive data applications and voice, you need to ensure that both applications perform as needed.
Echo cancellation: Due to the delay introduced by such processes as compression, decompression, packetizing and transmission, users may hear their own voice (an echo) across the frame relay network. This problem must also be dealt with by the public switched telephone network, or PSTN. VoFR equipment should have echo cancellation capabilities to prevent a noticeable echo.
Jitter buffer: Public frame relay networks are shared among many customers. No matter how much control you exert over your own traffic prioritization and fragmentation, you have no control over other customer traffic. This can lead to some of your voice packets experiencing 50-msec network delay, while others may get delayed during times of congestion. The jitter buffer temporarily holds incoming packets in order to assemble them in the correct order and recreate a high-quality voice signal. Like dynamic compression, some jitter buffers are also dynamic. This allows them to "sense" variable delay and provide a higher quality voice signal.
In the next issue, we will explain the areas where delay can be introduced when sending voice across the network.
RELATED LINKS
Dispelling the myths of VoFR
The newsletter referred to at left.
Voice over Frame Relay audio primer
Listen to a five-minute explanation of the technology.
GDC readies ATM, frame relay access offerings
Network World,
4/27/98.
Demands for speed and quality of service drive evolution of frame relay services
Network World, 2/9/98.
Voice over frame relay white paper from ACT Networks
Voice Over a Public Frame Relay Backbone -- Technological Considerations for Widescale Implementation
A paper from RAD
Data Communications.
Sign up for the NWFusion Focus: Frame Relay newsletter
Additional NWFusion Focus: Frame Relay back issues
