Dichotomy of Access and Backhaul Capacity Planning

By | July 20, 2011

I have listened to a few webinars recently by some of the major microwave backhaul vendors, all with the message that microwave has enough capacity to support required LTE data rates. It is evident that network operators have been pushing these vendors for higher data rates. Microwave after all cannot compete with fiber on capacity and MNOs (Mobile Network Operator) have been laying lots of fiber in anticipation of LTE network roll out.

I wondered as I listened whether MNOs use the same capacity metrics to plan their access networks as they appear to do on the backhaul network. If they ask for, say 675 Mbps for a backhaul link to a 2×20 MHz 3-sectored LTE cell, do they actually plan their access capacity using the 675 Mbps? If they do, the outcome will be lots of complaints from customers who cannot get their mobile applications to work because that inflated capacity number (675 Mbps) implies that the operator will run out of capacity on the site much sooner as it would have been loaded with a lot more customers than the actual capacity would allow (which should not be more than around 160 Mbps). On the other hand, the backhaul would be overprovisioned with excess capacity: there will be no performance problem there, but a lot of backhaul capacity sitting idle and much investment is tied up for perhaps years to come.

The effects of underestimating capacity on the access side result in the reverse situation: there will be ample capacity on the access side, but more investment is tied into access infrastructure since more cell sites need to be planned to support forecasted subscriber traffic growth. Backhaul performance will be poor. Another tricky situation!

What the above illustrates is that it’s important to have a good and measured account of what can actually be achieved in wireless networks. This is trickier with 4G LTE networks than with 2G/3G networks because LTE is packed with capacity-enhancing features that are very conditional. There’s no use in overestimating or underestimating capacity numbers. When ARPUs are stagnant and penetration rates at saturation levels, it’s ever more important to control the cost side of the equation. Having a firm grounding in what can be achieved, realistically and without the hype, is critical.