Speed alone is not enough – The challenges and opportunities for network managed latency

Speed alone is not enough – The challenges and opportunities for network managed latency

By Ken Ko, Managing Director at Broadband Forum

While speed is undeniably important aspect in broadband, there   is a continuing recognition that speed alone does not define a broadband subscriber user’s experience. This shift in focus among service providers on Quality of Experience (QoE), leads to reduced churn, new Average Revenue Per User (ARPU) growth opportunities, service differentiation and a decrease in customer support and network planning OPEX.

Understanding latency

A crucial factor in QoE is latency, recognized as one of the key user experience bottlenecks. Once a service provider can deliver higher bandwidth that is no longer a restraint on the applications demanded. Latency will become as important or potentially more important to the user experience.

Latency, an expression of how much time it takes for a data packet to travel from one designated point to another, heavily impacts on customer QoE and consumers are becoming more aware of this. Latency is impacted by a number of factors: the physical time it takes for the packet to travel from point A to point B; queuing caused by competing traffic; serialization, which is the time it takes to place the packet’s contents on the transmission channel; and packet processing time. Ideally, latency will be as close to zero as possible all the way to the user device. Too much latency leads to poor application behavior, causing consumers to become frustrated.

For service providers, this offers both challenges and opportunities. We as an industry need a way to explain the benefits of low latency in concise terms that are relatable to the average consumer in order to effectively monetize latency solutions. As an industry we need to create better metrics to help the end-user understand latency in terms of their specific application needs and experience.

The latency metric needs to be something more beneficial to a subscriber and operator than a one-off moment in time ping test.

Why does latency matter to the future of broadband providers?

Subscriber understanding of the importance of latency has been traditionally restricted to select “mass niche” application users, such as cloud gaming or industrial IoT. We are now entering a world where there will be new applications which have a broader appeal and expose more people to the discomfort of having poor latency. New applications are emerging to accommodate for the advent of Virtual Reality (VR), Augmented Reality (AR), and three-dimensional worlds with remote online interaction. With these applications, latency becomes absolutely critical to providing customers with an experience that feels realistic and comfortable.

Meta have indicated that this AR world is likely to require 10-20 ms of latency all the way to the device/headset. That is up to seven times lower latency than that required by most of today’s cloud-based gaming applications.

As mentioned earlier, some groups of customers are starting to feel the effects of latency and are ready to pay a premium for an improved service. One target group, labelled ‘gamers,’ have been pioneering this need as they are already educated and aware of the impact of poor latency to their user experience. Their gaming experience is directly affected by poor latency and for that reason, they are prepared to pay more or to move to a different service provider.

A recent Light Reading webinar poll indicated that over 50% of latency aware users are willing to pay an additional $5 to $10 more a month for the appropriate latency experience.

In addition to Gamers the COVID-19 pandemic introduced new customers to the effects of poor latency. Increased demand for multiple live 4k and 8k video streams, and the increasing popularity of VR headsets all have in impact on network latency requirements.

Similarly, homeworkers and businesses may well recognize the need to invest in premium broadband services that guarantee low latency video conferencing, webinar hosting and other business cloud-based applications.

Managing latency

Typical network quality metrics do not capture Application Outcome sufficiently. You can have identical bandwidth, average latency, jitter, and packet loss values, and yet still experience remarkably different quality with different applications. Application Outcome varies with how latency is distributed.

Broadband Forum’s Broadband QED (Quality of Experience Delivered) project, specified in TR-452, defines network quality metrics that capture variability in network quality, relating directly to Application Outcome and end-user QoE. The framework uses Quality Attenuation (written ∆Q) to address the need for improved performance measurements and analysis required by innovative broadband networks, tackling factors such as latency, consistency, predictability, and reliability.

Broadband QED provides a framework for decomposing a trip time into distinct components, matching them to the sources of performance degradation, for example, packet loss/delay, and relating them to geographical network topology, network features, and network load/scheduling. QED is a metric that builds a representative statistical distribution of latency. It measures the round trip times of a large volume of packets sent over a network segment over a certain time period or continuously. This makes QED a very capable metric for evaluating both latency and stability, and as such a suitable fit to address the latency challenge.

QED is the “game-changer” because it provides information to determine what the outcome of each application type would be. This ability to predict outcomes means we can apply state-of-the-art Machine Learning techniques to predict what the end-users’ Quality of Experience will be before it has happened.

Conclusion

There are both OPEX savings and commercial potential in offering prioritization and improved experience for certain applications as a Value-Added service. However, this should be targeted differently to each particular consumer group. This highlights the need to differentiate the offering of improved QoE and latency solutions initially to particular target groups and ultimately the wider Broadband subscriber market. Broadband QED provides the needed framework to specify, measure, and analyse the quality required for the applications driving these Value-Added services.