- Swallowing Yahoo may make Microsoft want to take a nap
- ISO approves OOXML as standard
- Maintaining integrity on the Net
- Microsoft caves, in part, to online computing
- Eyewitness to H-1B scammers
- Social networking hits the bar scene big screen
- Is the slow economy hurting high-tech sales?
- Take the smarts out of smartphones
- U.S. Immigration [USCIS] changes selection process for H-1B visas
- Will the iPhone force Apple to change course?
November 06, 2007 | Comments: (0)
Vetting the quality of VoIP
When it comes to measuring quality of service, experience is all that counts
Judging by the gamut of responses to my recent blog post "The lie that is Voice over IP," the verdict on VoIP is still out.
Sure, plenty are crowing the virtues of VoIP, but just as many are hanging up. And the fact that some users think VoIP is fantastic, while others have cancelled their services and are switching back to PSTN (public switched telephone networks), proves my point: VoIP quality is uneven.
In following up on that post, I picked up my landline and called Mike Hollier, CTO of Psytechnics, a company that measures the QoE (quality of experience) of VoIP and Video over IP rather than the more common QoS (quality of service).
The difference between these assessments is huge, and the fact that somebody is attempting to measure the quality of the user's experience represents a refreshing change from the typical analysis, which assures us that, technically speaking, the equipment is working flawlessly without taking the user into account.
It reminds me of the punch line of that old joke that says the operation was a success, but unfortunately, the patient died.
With deep roots in science as a lead researcher at British Telecom Labs for voice, video, and multimedia, Hollier worked on codecs and developed mathematical models of how the human senses work, all in an effort to better align the results of various measurement methodologies with how people describe their experiences of perceived phenomena.
So rather than just analyze the behavior of packets, Hollier's team looked at the waveforms of signals to see how humans would perceive them.
"We were trying to solve a whole family of perceptual engineering conundrums," Hollier says.
For example, on the visual side, BT Labs wanted to find a means for interpreting what the human eye sees and, by virtue of the boundaries of the information your eye can extract and your brain can interpret, use that as a basis for scientific measurements.
The results is a solution Psytechnics calls Experience Manager, which is a performance management solution for measuring voice and video capability over IP.
Voice has to be the first concern. Video, I suppose, will grow in importance once we get used to watching our favorite programs on a 2- or 3-inch screen.
Targeted at enterprise-level users, but not yet ready for major network providers at the IP trunk level, says Hollier, Experience Manager will give companies a tool that is lacking on the operational side of voice.
Current tools are not adept at recognizing poor voice performance. Network monitoring tools can't tell the network administrator that sounds are muffled or distorted or accompanied by an echo.
When Experience Manager identifies a problem, it sends alarms and alerts and troubleshoots the problem. The tools know the source and destination of each call and take measurements along the way to discern whether levels are "cold," a.k.a. too low, at a particular gateway, Hollier says.
According to Hollier, Experience Manager can even recognize a low-quality headset as the source of the problem or a low-quality sound card in the client.
What does this tell us about the state of the art of VoIP?
As is often the case with technology still in its infancy, the consumer becomes the beta test site on a grand scale.
The fact that Psytechnics thinks it alone has developed the tools to identify and provide relief to suffering VoIP users tells you that there is still a long way to go before VoIP comes close to offering the same quality provided by POTS lines.
This should also tell us something about the next hot high-tech space, UC (unified communications), touted by Avaya, IBM, Microsoft, Cisco, and Siemens, among others.
Think about this before you move forward on UC: Initially, companies will probably choose multivendor solutions. Each vendor will monitor and report and send alerts about their equipment, meaning, of course, when there is a problem the standard response will apply: "Oh, you're using our router with THEIR handset? Well, testing devices show us our equipment is working fine. The problem is with the other guys."
I'm sure over time there will be other vendors offering similar QoE solutions, but at the moment Psytechnics is the only game in town as far as I can tell.
What I like about Pystechnics' technology is that it looks across all vendor offerings, collecting end-point data and aggregating it for an overall view of video and voice performance. It is unique in that regard.
In the end, however, whether to buy into VoIP now is all up to you. When I was a kid, I kept asking my father when we were going to get a color TV. His standard response was, "When they improve the technology." Of course, I would have been happy with the technology as it was. After all, my friends all had a color TV.
You and your company are in sort of the same boat. Adopt an arguably flawed technology now or wait until they improve it?
Your call.
Posted by Ephraim Schwartz on November 6, 2007 03:00 AM
RATE THIS ARTICLE:
-
- COMMENTS
Yes, it's critical to measure "quality of experience," but it's still not just about lost packets, jitter and muffled sounds. More than 30% of the problems a VoIP user "experiences" are from outside the network: ambient noise, acoustic echo, mismatched speech levels, etc. Measure *that*! And the next challenge is to fix those problems. We actually do both (Ditech Networks) in real-time, on both ends of the call, and we do that for entire carrier networks, not just the enterprise. We've increased MOS scores up to nearly a full point, getting VoIP users to actually like and depend on the service for a change. When you *rely* on VoIP, quality goes straight to the top of the priority list.
Posted by: Ken at November 7, 2007 11:14 AMPeople need to see the difference between VoIP end-to-end and VoIP to the closet. I use VoIP phones in client offices to make my cabling easier, and to give them more features. Now, that VoIP phone is talking over the LAN to a PBX, which is then connected to a PRI T1 - so normal interfacing with the public.
VoIP over vast internet trunks is a gamble I don't take, but within a closed network, like an office, it's a great technology.
Posted by: ejoe at November 9, 2007 09:46 PMKen, you're absolutely right, you cannot manage VoIP quality of experience unless you're looking at issues introduced outside the immediate network ... noise, levels, echo etc. This is precisely what Psytechnics' Experience Manager brings to the party. It's the only tool on the market which can do this passively, non-intrusively, in real time. Most importantly, it will indicate where the noise or levels problems were introduced. It quite literally saves weeks of effort in troubleshooting VoIP networks.
Posted by: SH at November 12, 2007 05:12 AM
