This post is by Dr Michael Flanagan – CTO Mobility, JDSU
The presence of “extreme” data users is well-known in the industry and past reports have shown that as little as one per cent of the hungriest UMTS users consume half of all downlink data. The advent of LTE has introduced unprecedented downlink data rates in the wireless mass market. It has been postulated that the combination of extreme users and high data rates would result in a perfect storm where the hungriest users consume even more due to a high-speed delivery mechanism. Recent analyses confirm this prediction and show that LTE users are ten times more extreme than their UMTS predecessors.
The attached figure shows the LTE downlink data fraction as a function of user fraction for a Tier-1 market. In the upper-right corner we see that 100% of all users are required to account for 100% of all of the downlink data (as is naturally expected). In the lower-left corner we see that 0% of all users are required to account for 0% of the downlink data (also expected). If data consumption was uniform across all users, then we would expect a user fraction of 50% to correspond to 50% of the downlink data. But data consumption is far from uniform and a remarkable 0.1% of all LTE users are needed to account for half of the data.
To put this in perspective, imagine an office building with a thousand LTE users in it, and you line them up, not from tallest to shortest , but from hungriest user to least-hungry user. The first person in line consumes as much as the other 999 combined. The departure of this hungriest user means that total demand from the building drops in half. According to the market study, the first ten people in line consume 70% of all data.
LTE has introduced a new breed of extreme user. And this has important consequences for network evolution. The introduction of small cells where extreme users are located is an important strategy for responding to inexorably increasing data demands. This analysis of extreme users shows that a smaller fraction of small cells needs to be deployed to achieve the same relative capacity relief compared to prior UMTS-based estimates. As an example, a doubling of capacity is achieved with one-tenth as many small cells. While the number of small cells will still be large compared to the number of traditional macro cells, this result is good news for operators employing a customer-centric deployment strategy.