Increasing LTE's ability

Previously, in this series of Industry Insights on enabling LTE, I have addressed what is needed to backhaul LTE. I discussed how hybrid microwave radio could help meet those requirements. Specific LTE-proven backhaul techniques were outlined that can make LTE backhaul real. In many cases, operators need to maximise backhaul capacity for LTE.

Techniques LTE operators have considered to increase backhaul capacity include:

* Wider channels
* Multiple channels
* Adaptive coding and modulation (ACM)
* Higher-order Quadrature Amplitude Modulation (QAM)
* Multiple Input, Multiple Output (MIMO) antennas
* Ethernet and IP header compression
* Better IP planning

Some techniques to increase backhaul capacity are field-proven and accepted across the wireless industry. Other techniques are newer, not universally accepted and dependent on factors that may not be consistent in deployment. One approach is not technological as much as a realisation of how to use IP technology wisely.

Methods of increasing LTE backhaul capacity:

Wider channels
For LTE, wider microwave channels on the backhaul provide more capacity. For example, more water will flow through a bigger garden hose. At lower microwave frequencies of 6GHz-11GHz, channels are narrower due to heavy utilisation by operators. For wider channels, operators can shift their backhaul into higher bandwidths topping out at 80GHz.

Multiple channels
Another proven method of increasing backhaul capacity is using multiple parallel channels. Two or more links can carry portions of the same LTE transmission - a technique called link aggregation. Where a single large LTE data payload cannot go across one narrow microwave channel, link aggregation enables using multiple narrow channels until it fits. Multiple channels can also be enabled for LTE backhaul by using two antenna polarisations. Microwave technology now allows a signal to be sent along the vertical and horizontal axis of an antenna simultaneously, doubling the capacity.

Adaptive Coding and Modulation (ACM)
ACM maximises bandwidth through automatic adjustment of modulation and coding so the most data efficient (ie, highest possible) modulation is used given the atmospheric conditions. Adaptive modulation dynamically adjusts to virtually ensure maximum bandwidth - vital for LTE. Only under severe atmospheric conditions is modulation downgraded to ensure some connectivity, albeit at lower throughput. With adaptive coding, individual modulation rates optimise for maximum throughput or maximum signal strength - whichever is more important.

There is always the need to look at other technologies, including WiMax and WiFi, to see what approaches increase capacity and if they can be adapted to microwave. Current practices must also be examined to see if they can be extended to expand capacity where the payoff justifies the investment. Line-of-sight MIMO, higher-order QAMs and Ethernet and IP header compression are three current examples.

Higher-order QAMs
In microwave, usage of steadily increasing QAM modulations has enabled substantial capacity gains - from 8 QAM to 16 QAM to 256 QAM - current highest modulation in practical use. With LTE, some vendors have propagated 512 QAM, 1 024 QAM and 2 048 QAM as capacity solutions. However, for each higher step QAM, the capacity percentage gain is relatively less. From 256 QAM to 512 QAM capacity only increases 13%. By 2 048 QAM, percentage capacity increase is down to 10%.

Line-of-sight MIMO microwave
MIMO antennas have been common in WiMax and WiFi for years. The concept of using a series of 2x2 or 4x4 radios is not new. With enough separation, radios used in parallel on the same wireless link can double the capacity of the path. WiMax and WiFi terminals do not need to be separated much to make this practical. However, the distance needed to separate LOS MIMO microwave antennas on one tower is dependent on the frequency and link length, and would have to be at least 20 metres - too far to be feasible.

Ethernet and IP header compression
LTE is based on Ethernet, IP and the packet concept. Theoretically, as packets are sent across the backhaul, their headers may be able to be compressed, leaving more usable data capacity. Headers are compressed by replacing their data overhead with smaller placeholder bits before their packets are sent. When the packets are received on the other side, the full header is regenerated and reinserted. But only a limited number of packets can be manipulated at any one time, leaving this an incomplete solution.

Before LTE, cellular networks were based on time division multiplexing or circuit technology. Each mobile connection across the backhaul had to be maintained at a constant capacity, regardless of how much data was transmitted. The only way to add more capacity was to add more circuits.

There is always the need to look at other technologies.

However, with packet technology, if there's spare capacity on one path, it can be shared with another path that needs it. This is because IP transmissions are "bursty". They only require maximum capacity for rare instantaneous bursts. Otherwise, they are utilising the bandwidth at some lower level of activity. For example, three 200Mbps paths only require 400Mbps capacity, not the full 600Mbps, if they are "overbooked".

For mobile network operators moving into LTE, the watch phrase for microwave is "caveat emptor" or "let the buyer beware". There is no free lunch, silver bullet or panacea for the LTE backhaul bottleneck. If a technology solution to LTE backhaul sounds too good to be true, it probably is.

Always take outsize capacity claims with a grain of salt. Look at results from the field where the capacity improvement techniques have been deployed. If there are none or only laboratory-based results conducted under ideal operating conditions, be sceptical. As all African mobile operators know, the conditions in the field are often far from ideal.