Doug Simmons (13:18EST 1/4/14):
Hey Stephen, do you suppose that if New York City’s new mayor, of whom I’m not a fan, decided to impress me by decreeing that all carriers immediately deploy LTE Advanced throughout Manhattan and lift all data quotas and throttling, with carriers cooperating with each other to share all their spectrum with each other (and giving out replacement devices to all New Yorkers with radios that supported LTE Advanced on all these bands) — would that be good protocol and spectral efficiency to enable New Yorkers to do whatever it is they’re inclined to do on their cellular devices, unleashed, at praise-worthy performance?
Or is that simply out of reach without things like wifi availability and curbing cellular usage? And if it’s out of reach, do you think the math geniuses and engineers will ever be able to come up with a standard that would be efficient enough to handle a dense city like that?
Stephen Mesik (13:19EST 1/4/14):
I think it would be sustainable in the short term, but I think the larger, long-term issue is that mobile networks are like highways. The thinking at the time of the inception of the Interstate Highway System was that limited access and high speeds would satisfy demand for the foreseeable future, since capacity could be added simply by adding lanes.
What observation has taught us since then, is that demand exists because the presumed carrying capacity exists. In other words, as we add capacity more users are enticed to utilize and expect more from the service. Thus, a hypothetical twenty-lane superhighway through the heart of a major city would bear just as much traffic (as measured by the percentage of its total carrying capacity) as a four- or six-lane highway.
Likewise, I would anticipate that lifting bandwidth caps, throttles, and shaping on a LTE-A network with aggregated 100MHz carriers would simply lead to eventual saturation of its capacity, whether that limit occurs at the air interface or the lack of a sufficient backend network capable of carrying traffic to the NodeB. There will always be a bottleneck somewhere, and if the theory that drives (no pun intended) highway engineering today is at all applicable to mobile networks, the same result will occur.