Intervals to instants – The ultimate telecoms “game changer”

In order to achieve this goal, we need to quantify supply and demand, and do so in some common unit. Today we use “bandwidth” as a shared resource model. The nature of bandwidth is an average over an interval, as are related measures like jitter or loss rate.

Unfortunately, bandwidth is disconnected from the end user experience, and fails to capture the actual resource costs. Both of these are quality-centric measures, and there is no quality in averages over intervals. Hence bandwidth is an inadequate basis for trading resources under software control, since it satisfies neither side’s needs.

The ultimate “game changer” of telecoms is to adopt a new common resource model, so that “profitable” network resource trades can be identified and executed. This requires a new shared standard in which to quantify the supply and demand for information exchange. It must reflect QoE for the end user, and resource cost for the telecoms supplier.

This in turn means we must measure, model and manage the instantaneous network performance, since this is what the user experiences. The whole industry has to engage in a fundamental and unavoidable upgrade: from intervals to instants. This is the only possible way of capturing the essential quality properties.

The concept of quality attenuation (and the ∆Q calculus that quantifies it) is what makes this shift mathematically possible. You need a resource metric that “adds up” as an algebra, so you can do the kind of design and operational trade-off analysis that is taken for granted in every other engineering discipline. Nothing else does the job.

This move to a new instantaneous resource model and metric impacts people, process and technology. Each domain needs a “quality revolution” to an outcome-led approach. The scale of the change far exceeds what any one player in the telecoms and computing ecosystem can individually address. It is a systemic issue demanding a systemic response.

There is not only a new mathematics and science that requires dissemination and adoption. New measurement systems will be required to manage digital supply chains and to contract quality. New engineering techniques and processes are required to model and predict performance. New mechanisms can control quality at the finest timescales. These all have consequent OSS and BSS impacts across the whole ecosystem.

My sense is that this QoE-led networking is not just an R&D iteration of the current telco and equipment vendor business model, but rather is a new business (which subsumes the current one over time). That is because you cannot change a culture that rewards “more bandwidth with less QoE” into one that incentivises “less bandwidth for more QoE”.

You have to grow a new culture as well as new products. Salespeople used to being rewarded for more quantity will not overnight become ambassadors for a model of less quantity. Your products that are defined by quantity cannot (initially) sit in the same channel as ones where quality leads, since the marketing messages clash. The nature of the value proposition and revenue model change profoundly.

That places a huge onus on the leaders of telcos and their suppliers to locate and empower the “game changers” in their organisation. Thankfully, we have seen some early signs of promising activity, such as TMForum promoting the GC Index approach to locating these “obsessive creatives”.

I believe that the GC Index does for people what the ∆Q calculus does for instantaneous network performance. It is a transformative idea on what is the right “outcome-led metric space” for the feeling and thinking humans. The parallels are intriguing, and I would love for the telecoms industry to pioneer the application of GC Index in this related transformation.

In summary, telcos and their vendors need to engage with this unchanging external reality: the user’s application performance experience is dependent on instantaneous quality attenuation (and nothing else). As such, they need to begin to build the “pioneer” teams to explore the world of instants, and who are not encumbered by the historical baggage of intervals.

Without game changers, you will be left behind, playing the wrong game. Technologies like SDN appear to be solving the wrong problem really well. 5G is at serious risk of failure since its ambitions exceed the maturity of the engineering tools on offer. There is a serious risk of collapse back to monopoly and vertical integration if we do not solve these core quality and science issues.

Survival is not mandatory, and I expect the cloud and data centre players to expand their activities into the network over time. They only need to modify their existing demand-led systems. These were built for a compact and regular design space of data centres, and need to be modified to work in a distributed and irregular one of wide-area networks.

Amazon, Google and Facebook could easily wipe out a lot of the value of the telecoms equipment vendors, and do it very soon. They would then be in a position to suck out the value of the capital assets of the telcos. This would not be good news for a lot of pension fund investors expecting future steady dividends.

I believe a necessary and early step is to initiate an institutional response to the “instantaneous quality engineering gap”. Some kind of organisation is required to formulate the right questions, initiate the right standards, disseminate the basic enabling R&D tools, and manage the common IPR (as has happened with 3GPP, for example)

This requires someone of the appropriate credibility, interest and connections to head up this industry initiative (and that isn’t me!). Any ideas who might be “game for a change”?