Last year, many operators achieved respectable penetrations levels for public switched telephone network (PSTN) quality plain old telephone service (POTS) and this year will be looking at new offerings to continue expanding their telephony base. Converged services and multimedia telephony are part of that strategy, and providing these offerings raises a bunch of new issues.
Let’s begin with simple quality measurement. I spoke about this topic with Benjamin Ellis, Psytechnics director of global marketing, who provided some insights into the necessary direction of quality assurance for voice in a multimedia world. Psytechnics is one of the root companies in voice quality measurement and was the first company to create a product based upon quality testing done with British Telecom in the late 1990s.
Ellis began by saying, “Over the past five years, voice quality measurement has gone from a network edge measurement to an end-to-end metric.” He then explained why this makes a simple mean opinion score (MOS) insufficient to gauge quality.
“Even though the quality of an IP call in cable company networks is excellent, from an end-to-end perspective, voice quality in phone calls has actually declined over the past six years. This is due to the many networks involved in an end-to-end call. A VoIP (voice over Internet protocol) call from your network, for example, typically coexists with TDM (time division multiplexing) and cellular technology.” There are several points in the network where call quality can degrade, and consistent measurement has to be made at each of them to ensure quality is maintained.
Gateway and end point measurements provide an indication of the quality of a voice waveform as it enters or leaves a particular network, but by themselves do not provide enough information to determine causes of degradation. IP packet loss, for example, may be one cause, but there are other conditions that can be just as detrimental. Even something as simple as the differences in phone receivers contributes to call quality degradation.
Compounding the problem is the fact that not all measurements of call quality are the same, although they may have the same name. “A MOS score, for example, can be determined by several methods,” Ellis explained. “The original gold standard was a weighted average derived from a panel of listeners, and even that is very subjective. Today, however, many MOS scores are derived from indirect measurements, such as packet loss, which can be measured differently.” When a network manager needs to assess what happens to voice quality as a call traverses a network, it is therefore necessary to compare apples to apples. A good MOS score from another provider, even though apparently quantified as a number, may not mean the same as your MOS score in terms of listening quality. Even the methodology of obtaining the score can be different. For example, PESQ (perceptual evaluation of speech quality), one of the more widely used measurement techniques, produces a MOS using an injected signal, rather than an actual call. Possible solutions Fortunately, there are solutions to these challenges. Adherence to new standards defined by the International Telecommunications Union plays a major role. The P.563 standard for speech quality measurement has been around for some time now, but it has only recently been complemented by the newly ratified P.564 standard that specifies how to monitor IP packets. Both need to be combined in call quality measurement, along with factors such as background noise and conversational clarity. Ellis termed such monitoring “Experience Management,” the concept used in the Psytechnics Speech Monitor (PSM) product.
Ellis predicts that in a multimedia world, overall experience will become just as important as features in determining subscriber loyalty. He noted the trend toward wideband codecs, particularly in applications involving both voice and video. “A wideband voice codec has a 16 kHz sampling rate, double that of the conventional codec used in typical PSTN applications,” he explains. “The result is conversational quality that is the same as being in the room with the other person. In the future, this difference in quality will determine choice of a voice service provider.”
Ellis may be right about the importance of voice quality to subscriber experience, but there will be other factors related to content that will also be important. Adding SMS Readers of this column will recall that I have suggested the value of adding simple messaging system (SMS) technology to cable telephony offerings. This feature is extremely popular with cellular users and has even generated a sort of language shorthand for conveying email type messages on phone displays. Last month, I noted that Integra5 was looking at adding SMS applications as another offering tied to the integration of TV and phone. With cable’s interest in the quad play, it’s possible we may be offering SMS as both a cellular and video service.
Opening the door to text messages, however, can also provide an entry point for SMS spam. Vince Kadar, CTO of Airwide Solutions, points out how this occurs.
“SMS uses the SS7 (Signaling System 7) network to deliver text messaging. A problem can occur when the SS7 network is ‘hijacked’ from a foreign gateway by a party spoofing the network as a legitimate device. The spoofer can push a large number of messages to phones in your network. A common scam is to send a message stating the called party has won something and should call back.”
Cellular roaming is the loophole that provides access to SS7. When a subscriber roams to another network, that network is responsible for validating the roaming phone and forwarding messages only from legitimate roamers. Although there is a process for validating legitimate roamers, according to Kadar, not all cellular providers outside the United States use it, opening the door to spammers. The problem carries monetary implications, since subscribers are charged for the spoofed call, as well as for calls to the number in the spammed message.
Airwide Solutions offers a fix to the problem in the form of a network element that proxies all traffic and validates all messages at the interface between the mobile switching center and its subscriber location register. The solution includes the capability to create SMS blacklists and whitelists to specify which sources are acceptable for SMS messages. Screening of this nature is beneficial to a cellular subscriber, but would also be an asset to a cable operator if SMS becomes part of the video experience. (Imagine subscriber reaction if unauthorized outside parties could add crawl messages to video programming.) Big ol’ world As this discussion illustrates, it’s a bigger world when technologies converge, and learning more about the components of converged networks outside our normal universe can be helpful. Along those lines, I recently paged through a new book, Wi-Fi Telephony by Praphul Chandra and David Lide. This text is a relatively easy read and contains a high-level look at not only Wi-Fi networks, but also cellular technology. Although it centers on global system for mobile communications (GSM) and is light on code division multiple access (CDMA)-based 2G and 3G mobile technology, it is still a good reference for anyone looking into fixed-mobile convergence (FMC). Justin J. Junkus is president of KnowledgeLink and telephony editor for Communications Technology. Reach him at firstname.lastname@example.org.