SDN primed to slash costs on the mobile backhaul network
And when they do adopt, it's even more certain that it will be in specific areas rather than as a blanket approach. One area that might benefit greatly from the SDN treatment, say Strategy Analytics and Tellabs, is mobile backhaul.
The pair have been looking hard at backhaul economics and how to optimise them. In previous analysis they uncovered what they termed a 'backhaul gap' between the projected rise in mobile broadband demand and the backhaul spending earmarked to cater for the growth.
If operators intended to maintain the same quality of experience, the analysis pointed out, they would essentially have to spend an extra $9.2 billion in the backhaul network to ensure it delivered QoE under the much higher loads. What's more, the gap analysis had already factored in the arrival of dramatically improved backhaul transport technology.
What's the solution? Well you knew all along: Software Defined Network' (SDN) investment.
The full report - available for downloading - is The Role of Software Defined Networking (SDN) in Bridging the Backhaul Gap.
It doesn't close the gap entirely on these calculations but it does get nearly halfway there by automating processes, squeezing down on OPEX, and by optimising resource allocation... it's amazing what a bit of clever software can do which, to be fair, is what many of our interviewees have been saying for a year or so.
SDN is a re-run of the PC
"The last time we saw this
magnitude of a shift was at
the end of the mainframe"
The Tellabs report calculation is that if carriers implement SDN in the backhaul network they can reduce that backhaul funding gap to just under $5 billion by 2017, down from the $9 billion posited before.
Strategy Analytics looked at three key applications.
Cloud-RAN (C-RAN): this is where you use "fronthaul" - connecting remote radio antennas and radio heads directly over fibre to base stations where coding and decoding takes place.
Small cell deployment: these are designed to meet peak demand so clusters can be dynamically powered up and down and linked over multiple paths.
Metro Aggregation/Load Redistribution: exploits partial mesh or ring connectivity to improve bandwidth utilisation and ensure end-to-end performance for high value traffic.
Local breakout: SDN can make intelligent decisions to get some traffic onto the Web as soon as possible, rather than sending it through the core.
WiFi offload/Video redirect: again putting some traffic 'seamlessly' onto WiFi when the going gets tough.
These applications and doubtless many more, can be put into operation to dynamically refine the backhaul network and thus save on both OPEX and CAPEX it's claimed.