LATAM Regionalization

A high-volume customer in Latin America cut end-to-end round-trip time (RTT) from 191.3 ms to 99.5 ms after we moved their calls to a LATAM point of presence (PoP). The same setup is available for other regions, as described in Regional deployment.

Call quality defines each metric used below.

Background

The customer handled high concurrent call volumes in Latin America, and its calls egressed through US infrastructure. During peak hours, call quality varied.

Problem

Calls had higher latency, packet loss, and packet delay variation (PDV, also called jitter). Callers heard choppy or robotic speech, and occasionally a brief squeal when packets arrived in bursts.

Solution

• We deployed a point of presence in LATAM, so media and signaling run close to the callers.
• We routed traffic that still crosses regions over the Synthflow backbone instead of the public internet.

Shorter paths lower RTT and one-way latency. The backbone avoids congested public segments, which reduces hop count, jitter, and packet loss under peak load.

Results

Results in your region depend on last-mile access and your carrier.

Methodology

• RTT and path: mtr -u -c 200 <media-endpoint> and ping -c 200 <media-endpoint>.
• RTP stats: a sample of at least 100 calls captured in Wireshark, with RTP Stream analysis for jitter, loss, and skew.
• Synthetic traffic: sipp for call timing, and iperf3 -u for UDP jitter and loss when a test server is available.
• MOS: computed with the E-Model from measured one-way delay, jitter, and loss.