However, the validity of the study has been questioned by some industry watchers who claim Brix’s technology cannot conclusively determine actual call quality. Indeed, many in the industry argue VoIP calls have incrementally increased in quality recently.
I wouldn’t say I draw the same conclusion as the report has, said IDC research analyst for VoIP services Rebecca Swensen.
With the technology that’s being coming out recently, there’s been an increase in the quality of service, whether it’s through latency or jitter, over the past 18 months, Swensen said.
Brix collected its data using its TestYourVoIP.com free testing portal that measures the quality of consumers’ broadband and Internet phone connections. Having collated the results for the past 18 months, Brix said the portal showed nearly 20% of calls were of unacceptable quality.
The study did not specify what types of quality issues were most to blame, but Brix measured signaling level metrics, such as the time it takes to get a ring tone, media level metrics, such as voice quality, and network level metrics, including packet loss, jitter and latency.
Brix CTO Kaynam Hedayat said the company could monitor QoS in three ways: by generating synthetic calls into the network; monitoring live user calls on the network; and collecting information tracked by VoIP devices in the network.
However, the company used only the first method, generating synthetic calls, for the study. It calls this method the Brix Tri-Q Analysis.
Synthetic calls are made through a Java agent from the user’s desktop to any seven locations across the globe, Hedayat said.
But Swensen said this method essentially is testing broadband delivery capability for VoIP calls and not necessarily VoIP calls.
Hedayat countered by saying Brix’s tests established VoIP calls by using signaling and media. Through these calls we are measuring application level metrics … and not the broadband capability, a network level metric, he said.
Furthermore, the metrics that we provide are end-to-end metrics for VoIP calls that include both broadband and core part of the Internet.
Swensen and others were not convinced. I haven’t read too much into [Brix’s report], Swensen said.
However, she agrees that VoIP QoS will be an issue at some future point when VoIP adoption rates pick up.
But Brix asserts voice traffic already is competing with different kinds of broadband traffic that are generally more bandwidth hungry than voice, such as video and music downloads.
Our results clearly show that VoIP quality is degrading and it is our belief that the degradation is due to lack of differentiation of voice streams, Hedayat said.
The initial results from Brix’s survey were based on calls made from all over the world. The company is working on a second phase of the study, which it expects to complete by the end of the year, which will provide more detail based on geography and service provider, Hedayat said.
In the same announcement, Chelmsford, Massachusetts-based Brix also said TestYourVoIP.com would be available as a Google Gadget mini desktop application.