When does latency become an issue to the flow of voice traffic?

When does latency become an issue to the flow of voice traffic?

We are decommissioning Newbridge MUXs and tie lines and redirecting voice traffic from existing PBXs to a Cisco VG200 and across leased lines (T1 or E1) that incorporate existing Expand compressors and Cylink encryptors on both ends of circuits. The edge routers use LLQ for QoS. My concern is when does latency become an issue to the flow of voice traffic as these packets traverse the WAN.

I believe the amount and ratio of data to voice traffic will determine when latency starts to become an issue. Currently, there is an average 250ms response time (ping test) between the States and a remote branch in Argentina. The compressors have caching mechanisms to assist with traffic flow and may employ QoS mechanisms. Is this a recommended design? What are our alternatives?

Your thoughts are appreciated.
Rick

    Requires Free Membership to View

    Login

    By submitting your registration information to SearchNetworking.com you agree to receive email communications from TechTarget and TechTarget partners. We encourage you to read our Privacy Policy which contains important disclosures about how we collect and use your registration and other information. If you reside outside of the United States, by submitting this registration information you consent to having your personal data transferred to and processed in the United States. Your use of SearchNetworking.com is governed by our Terms of Use. You may contact us at webmaster@TechTarget.com.


The amount of data and voice traversing your private lines (for WAN connectivity) will definitely affect the quality of the voice. The primary performance metric you should be looking at is jitter. Even if the latency may be high (not unreasonably high though), if the jitter is good the quality of the voice will be good. An alternative you may want to consider is running some or all of the traffic over an IP VPN. Most public IP networks have immense capacity so you're likely to have no problems delivering great quality for applications over them. You may have to prioritize traffic on the last mile circuit if the utilization is high.

This was first published in September 2002