Putting LTE and 4G into Perspective

Peak data rates capture headlines. LTE’s current promise for 2x2 MIMO is a download speed of around 180 Mbps and when (if) the evolution of this technology LTE-Advanced measures up to ITU’s 4G requirements (IMT-Advanced) we’re supposed to get 100 Mbps for high mobility and 1Gbps for low. But don’t hold your breath. The gap between peak and average data rates for all variants of cellular wireless is huge and in the case of 4G it could even be 100:1. More on this in a moment.

That gap was one of the reasons why 3G didn’t live up to its early, over-hyped expectations. In this case the requirements (IMT-2000) didn’t say anything about capacity — the number of users in a call — or coverage — performance levels at the edge of the cell. Instead the focus was on the single-user, peak rate of 2 Mbps. IMT-Advanced does address these issues; it sets targets for average spectral efficiency and cell edge performance, but if operators focus more on the peak rates rather than the averages then the expectations for 4G will outstrip reality. In addition, 3GPP has emphasized that average and cell edge targets are more important than the peaks.

System capacity is another important issue. Today it’s mainly consumed by voice and text messaging, but much more capacity is needed in order to accommodate the higher rate 3G, 3.5G and 4G data services. Capacity per unit area is determined by a combination of spectrum, spectrum efficiency and the number of cells. If there is insufficient capacity then the peak rates cannot be achieved and right now the gap between peak and average rates is about 10x,and by 2015 peak rates will have outpaced system capacity by 100x unless action is taken.

These figures as well as some of the background explanation in this short article come from an excellent paper published in the Agilent Technologies Measurement Journal (“IMT-Advanced: 4G Wireless Takes Shape an Olympic Year”: http://cp.literature.agilent.com/litweb/pdf/5989-9793EN.pdf).

Additional capacity will come via additional spectrum and improvements in efficiency, but there are limits. Spectrum is an international political issue and we might only expect a 2x growth by 2015. The targets for efficiency at around 3x are very challenging, so any big boost has to come from the third axis: the addition of many more smaller cells. Current deployment is centrally managed and it cannot scale to the expected demand (10x rising to 100x). The answer to smaller cells has to be femtocells in the home and that fact has been recognized. In April the Femto Forum, 3GPP and the Broadband Forum announced that the world's first femtocell standard has been published by 3GPP (http://fixed-mobile-convergence.tmcnet.com/topics/mobile-communications/articles/53785-first-femtocell-standard-published-3gpp.htm).

Agilent’s paper gives details of a Qualcomm simulation designed to show the difference between the capacity of a single macro cell versus the higher median data rates and capacity of adding femtocells. A single macrocell with 34 distributed users and peak single user data rate of 17 Mbps has a median data rate of just 40 kbps. A 3x improvement could be realized with new higher efficiency systems that would raise that figure to 120 kbps, but it would be expensive. With 24 active femtocells in the macrocell it is possible to realize an improvement in capacity that would result in a median data rate of 8 Mbps, some 200x higher than the macrocell on its own.