Summary
Shannon’s limit has been an ongoing concern in telecom for decades.
Analysis
Shannon’s limit is not exactly an earth shattering concept. Funny thing is it is the simple ideas/concepts that often are the most profound. The idea is so easy to understand that up until recently most transmission engineers, RF engineers, and telecom/IT engineers simply ignored it because there had always been a technology solution or a technical implementation that could mitigate bandwidth limitations and transmission errors.
The reality is that Shannon’s limit is simply brilliant in its simplicity and its profoundness.
To date, engineers have had the luxury of never ever having to deal with the error or bandwidth issues raised by Shannon because the industry always had a bigger transmission medium to use or an engineering technique to mitigate errors and bandwidth limitations. An example of technology mining capacity out of radio would be digital versus analog. An example of an engineering technique to mine more capacity in a cell site would be sectorized cells versus the original single sector cell sites we used to put up.
Transmission errors have always been an issue. In the old days of voice only, interference simply either made the connection noisy or degraded the quality of the connection to the point where one party could not recognize the voice of the other party. However, as soon as non-voice data was being transmitted missing bits became an issue. Suddenly we are faced with the Internet and then suddenly we are faced with wireless Internet access. The last 20 years has been about pushing more data into a finite sized pipe without losing any bits. Compression algorithms have improved. Protocols have been designed to retransmit bits that have been determined missing.
What do femtocells bring to the table? In my opinion the beauty of the femtocell is that it brings network intelligence and capability closer to the user. Access point base stations can be used on fixed wireless environments and mobile environments. Theoretically they can increase capacity without interference. This reminds of the smart base stations the industry was trying to design back in the early 1990s but back then the chips to make it work did not exist. It all sounds very impressive. Is interference a legitimate concern? Yes, but there are ways to mitigate the interference; albeit not very pretty and will certainly upset some consumers.
However, what we really need to be concerned with is not interference but something completely different. What we are really looking at in the case of femtocells is a network architecture that would push existing cellular architectures to the limits of their design. Bottom line the cellular network as designed today involves a fabric that enables core switching entities to manage the traffic of a couple of thousand cell site within a given geographic area. Femtocells basically increase the number of cell sites in the same coverage are to a few million cell sites. The carrier has just gone from a big network to a show stopper and worse still all of these femtocells are using the same frequencies as the bigger cell sites. Let’s assume you can mitigate the interference issues by using more spectrum, self reconfigurable cell site technology, or simply better bandwidth management. The current underlying switching fabric cannot handle femtocell site deployment configuration I am describing. In a worse case scenario the switching fabric might come to a grinding halt. Think 5 pounds of stuff being jammed into a 10 pound bag. Enough to scare the heck out of any engineer having to build and operate the network.
I am still a supporter of femtocells because of the technology’s potentially beneficial impact on mobility and fixed wireless. Femtocells are not THE solution but they are a part of the solution. The best engineering solutions to problems are compilations of many smaller solutions.
