Three Views of Quality of Experience

James P. Cavanagh

From the late 1700s to the early 1800s, the Japanese wood block artist Katsushika Hokusai created forty six wood block prints and, for reasons understood only by the artist, entitled the collection “Thirty Six Views of Fujiyama”. Among the most recognizable, shown at left, is Kanagawa-oki nami-ura, The Great Wave off Kanagawa.

The thing that strikes me most about the Thirty Six views series is that there is a single unifying theme, Fujiyama (Mount Fuji), but Fujiyama’s role is ever-changing. At times Fujiyama looms large, as in Rainstorm Beneath the Summit, while in others, such as Great Wave off Kanagawa, Fujiyama is a minor element, diminutive though still a powerful part of the picture. So, it is, with Three views of Quality of Experience in which I will present four different perspectives on QoE, all different, yet all with a unifying theme and varying emphasis.

Definition: What is QoE?

QoE is quality of (user) experience. QoE is a touchy-feely qualitative measure that is very fragile and subject to change, very hard to pin down exactly. Some might even call it fleeting or ephemeral. QoE is subjective. It is subject to the mood, gender, age, prior experience, expectations, environment and present health of the listener and can change more than a little bit from measurement to measurement. It is more of a feeling than anything else: the type of a thing engineers hate most but the opinions to which users cling tenaciously. QoE is quite difficult to describe in a trouble ticket and even more difficult to quantify in a Service Level Agreement.

QoS, on the other hand, is qualitative: it can be measured or calculated precisely. QoS is objective, undistorted by personal opinion or mood. QoS can be measured repeatedly and consistently under the same conditions. How many packets lost per hundred transmitted? How much inter-packet delay? How much variation in the delay? These are components of QoS.

View 1: Service Provider’s View of QoE

Organizations providing Voice over Packet (VoP) services have struggled, and are still struggling, to provide a service that replicates historical Plain Ol’ Telephone Service (POTS) over a groaning IP infrastructure that was not built to handle a global multimedia load of data, telephony and video. Today’s services are predominantly VoIP-based and most recently are delivered using the Session Initiation Protocol (SIP). Going into 2008 it is a bit hard to believe that the era of IP Telephony started 14 years ago and that all services are achieving a fairly consistent level of service albeit not quite yet at the level of consistency, quality or reliability of POTS.

The question now is “what is next for VoIP service providers?” With reasonable consistency in quality, reliability and pricing, the next era will be the era when VoIP must surpass POTS’s quality of experience perceived by the users. One of the areas of positive differentiation, addressed in a previous Eogogics ezine article, is wideband voice wherein more of the actual voice quality and meaning are delivered. Wideband voice is one of the dimensions, including traditional QoS areas such as tighter control over packet loss, delay, delay variation and availability and including new considerations such as time-to-dial-tone (TTDT) and other metrics that influence the ultimate opinion of the user.

It is also noteworthy that current packet voice service providers are all over the map in terms of QoE but none are where they ultimately need to be to be successful in the long term. It is also noteworthy that in the service provider picture QoE is almost imperceptible in the background with price and availability looming larger in the foreground.

View 2: Manufacturer’s View of QoE

As service providers begin to wake up and understand the value of QoE, it is the job of manufacturers of the equipment used in service provider and large end-user systems to incorporate elements that allow QoE-enhanced services to be delivered. Big bandwidth alone will not do the job. Dynamic buffer handling, prioritization, queue management, feedback loops, wideband codecs and a plethora of other tools must be provided to allow service providers, and large end-user organizations acting as internal service providers, to deliver QoE-optimized services that surpass POTS.

In the manufacturer’s view QoE is a larger element but is still not the central theme though that is changing as manufacturers vie for positive differentiation and the resulting market predominance and carve out the market niches which they may occupy for decades to come.

View 3: VoP Users’ View of QoE

More than one VoIP initiative has suffered an early death due to user revolt based on nothing more than “it doesn’t feel the same”. The objective, of course, in many cases today is that the new system feels the same as the old system, but all that is changing. With QoE-enhanced systems, the new system should feel better than the system it is replacing.

Voice, and, in fact, video QoE is hard to define but users generally recognize it when they experience it. And then they want more of it. With all, or most, other factors being the same QoE, that “little something extra”, will increasingly become the differentiator, the tie-breaker that determines which system is implemented and which one is sold for its scrap metal value.

Another odd phenomenon, from the user’s perspective is that somewhere during the 1980s, telephony users in developed nations began to give up high-availability, consistently “toll quality” phone service for very low availability and inconsistent mobile and, eventually, cellular phones. QoE represents a return to an emphasis on the quality metrics long held near and dear: always-ready service that works consistently and delivers a high quality audio experience. From the user’s QoE perspective, in fact, the high quality audio experience covers not just the live, interactive human speech of telephony but also the fact that the device with which they are interacting produces music and also sound tracks to accompany their increasingly mobile/wireless video experience, as well.

From the perspective of the user QoE is the picture, literally, and the sound. Many older users long for the simplicity of the old desk phone: a simple hardwired device with 10 digits, a # sign, a * and maybe, just maybe, a hold button and some multi-line buttons that light up if you have more than one line. Today’s user, however, would not recognize one of these devices if it showed up out of the clear blue sky. They opt for small hand-held computers with complex user interfaces and often full keyboards, but are setting ever higher bars for their service, be it landline or untethered.

View 4: QoE as Objective/QoS as Tool

And, in a manner consistent with Hokusai’s math, that brings us to the fourth of our three perspectives. QoS has long been the domain of the network engineer and many of us — because this author, too, is a network engineer — fear the day when QoE will take over and replace QoS. But we don’t need to worry. In fact, QoE will never replace QoS. We need both. Quality of (user) Experience is the objective: it is how our work will be, and is being, judged while Quality of Service is how we achieve QoE. We need both and, in fact, both will be with us until the end.

To be successful we engineers will have to learn to translate. We will have to learn how to translate those metrics that we hold near and dear, such as packet loss, delay and delay variation into squishier, more subjective measures such as calculated Mean Opinion Score values that are not exactly 4.21 of something, but rather 4.2 +/ .4: less precise but more meaningful. We must be able to translate into “good”, “better”, “best” or “telephony quality” vs “CD quality” and similar units that will press the right buttons in the heads of our users and let them understand what all of our behind-the-scenes work means to them.

For engineers, QoE is a central figure in the picture and getting bigger, but our traditional elements are all still there, albeit ones that require a close look at the picture to truly appreciate.

Conclusion

QoE started as a video consideration and soon grew to include voice and now has been wrapped around the entire multimedia experience that is increasingly being called unified communications. QoE now knows no bounds: the concept applies equally to wireline and wireless, to data, voice and video, to “phone calls”, video conferences and telepresence sessions and to the home “TV” experience. QoE is all around us and must be considered seriously by service providers, manufacturers, end-users and the engineers who support these groups and their users. QoE has become a permanent, and important, part of the picture.

Editor’s Note: Jim, author of half a dozen books on telecom, has 35 years of hands-on experience in data networking and optical technologies. A dynamic and entertaining presenter, he teaches many of our courses on IP and MPLS, IMS, SIP, VoIP, SS7, SNMP, SONET/SDH and DWDM.  Read bio.