When we decide on an internet service provider we mainly care about monthly price and service speed. For example, in the US an internet service provider offers 150 Mbps for $55/month and 250 Mbps for $69/month. The more you pay, the faster service you’ll get. This is not necessarily true for the mobile internet service as the speed of wireless cellular service cannot be guaranteed. The speed depends on how far a user is from the cell tower, indoors or outdoors, wireless spectrum bandwidth, and how many other users are sharing the service at the same time. The wireless internet speed clearly matters to consumers and many of us have seen TV commercials talking about how one cellular network is better than another. Companies such as RootMetrics, Ookla and others have also made it easy to understand and compare mobile network performance. In this blog, we will discuss the factors affecting wireless internet speeds: signal quality, wireless spectrum bandwidth and mobile traffic demand or load, using a simple analogy and then extend the discussion to show how a mobile network operator can improve network performance and measure the improvement using the existing Key Performance Indicators (KPI).
Let us look at these three factors affecting the wireless internet speeds. First, consider the number of users accessing the service. It is intuitively easy to understand that the more users are using a service, the slower it gets. A simple analogy is a highway getting jammed when more commuters are rushing to and from work during the peak hours. There is not much a mobile operator is able to do to reduce the traffic, and arguably it should not attempt to do it under normal circumstances. So, the question is what can they do?
In the theory of communications, information rate depends on the bandwidth of the transmission medium and the quality of the transmitted signal, defined as signal-to-noise power ratio. The Shannon-Hartley theorem sets an upper bound for the channel capacity, C as:
where B is channel bandwidth and S/N is the signal-to-noise ratio.
In the previous example of a highway, the analogy for the bandwidth would correspond to the width of the road. Increasing channel capacity by increasing bandwidth is equivalent to making the road wider by increasing number of lanes. Now in reality of course, it is not that simple to build wider roads with more lanes, just as it is not that simple to increase the bandwidth since the wireless spectrum is a scarce and often expensive resource.
Since the mobile network operators do not have unlimited frequency spectrum they are concerned about how well the spectrum assets are utilized. This is often measured as spectral efficiency, the channels ability to carry information given fixed bandwidth.
In an expanded definition it refers to “the information rate that can be transmitted over a given bandwidth in a specific communication system. It is a measure of how efficiently a limited frequency spectrum is utilized by the physical layer protocol.” From the above Shannon formula for channel capacity, spectral efficiency can be numerically stated as C/B. The value is expressed in bits per second per hertz (bps/Hz).
Using the previous simple analogy of a highway, if we wanted to enhance the capacity or throughput of the highway without building more lanes we could let the commuters use buses instead of private cars. In this way the highway carries more traffic in terms of the number of commuters, and the efficiency is higher. Similarly, spectral efficiency of a communication system can be enhanced by packing more information, bits, in a single transmission. This, however cannot be done unless the quality of the channel is sufficient, just as a bus cannot access a road that is not well paved.
This brings us to the third factor that influences wireless speeds and network performance. The last term of the Shannon-Hartley theorem is the signal-to-noise ratio or simply signal quality. Improving signal-to-noise ratio will improve the channel capacity and spectral efficiency. In an ideal world, the cell towers could be placed uniformly giving high quality signals throughout the surrounding coverage area. In practice mobile network operators are limited to certain tower locations and they try to improve the signal quality by extensive radio network tuning and optimization work. The lack of freedom in choosing site locations and building new cell sites also limits the operators’ ability to increase capacity in dense areas. And that is exactly where Blue Danube’s Massive MIMO solution can help. Focusing radio signal to where the users are and steering clear from interference, the solution improves signal quality and spectral efficiency. In addition, serving users with different beams will improve the overall system capacity and user experience.
Now, let’s see how spectral efficiency can be quantified through KPI’s. Since the unit of spectral efficiency is bps/Hz, one could imagine that the numerator involves throughput and the denominator is the bandwidth. To also account for how much of the bandwidth is being used, one should include a measure of utilization in the denominator (in other words, a fully utilized channel that carries the same amount of traffic in a fixed time would be less ‘efficient’ than a channel that is able to carry the same amount of traffic in the same timeframe while only occupying the channel half as much).
Let’s now gather the KPI’s components for the spectral efficiency formula:
Channel throughput can be calculated from data volume (in bits) transferred over a period of time. With that, the components for calculating LTE network spectral efficiency from the KPI’s:
The formula for hourly spectral efficiency then becomes:
Similarly, to get the daily spectral efficiency, one can simply add the total data volume throughout the day and divide by the aggregate utilization for the same 24 hours:
Spectral efficiency is a good, quantifiable and comparable metric characterizing the performance of a wireless network. Getting familiar with this metric can help mobile network operators properly monitor the efficiency of their most valuable network assets.