The Flat Mobile Network

Mobile data traffic will outgrow processor performance, not just ARPU. Network architectures with multiple layers which process the same data at different nodes in the hierarchy will fail to keep up.

Network technologies with highly capable but deep protocol stacks that process the same data with different protocols in the stack within a node will also fail to keep up.

A paradigm shift in networking is necessary to arrive at an architecture that will scale by design:

• Only two layers in a mobile network should touch mobile data and that too with the shallowest protocol stacks for the least cost - the packet to wireless modem layer at the edge and the packet routing layer above
• For the packet to wireless layer, wireless and routing functions should be etched into the same modem silicon
• Growth in network capacity should be achieved by replicating silicon and densifying nodes, not continuous change in algorithms and architectures - a process that detracts from the business case for developing silicon
• Passive optical data transport between nodes will become indispensable
• Passive optical routing across the network will complete the transformation

Apart from reducing CAPEX and OPEX, such an architecture would consume the least power per Hz, bit or FLOPS. The planet and its people will stay cooler and breathe better.

Companies catalyzing this revolution won't be disruptors out there to win mind-share, the technology leadership mantle, venture capital or market capitalization, rather they'd be ensuring their survival. They'd just win all those other things along the way.

References

• “Exponential Laws of Computing Growth”, Denning P. J., Lewis T. G., Communications of the ACM, Vol. 60 No. 1, Pages 54-65
• “Annual Year-End 2016 Top-Line Survey Results”, CTIA, https://www.ctia.org/industry-data/ctia-annual-wireless-industry-survey