The future billboard Working on many things at the same time, the tunnel vision sets in, so every once in a while, I try to take a I step back and reflect on where we are in LTE land.

What I see is a early market that has yet to find it’s purpose in life. Yeah yeah, consumers are buying devices like Samsung Galaxy III/IV and Apple iPhones that are LTE enabled at a rapid clip in the US, but at this point the device is driving the sale not the LTE itself.  

There have been a couple of interesting developments, such as Audi and GM deciding to leverage LTE for in car data links and even going as far as choosing to use VoLTE for voice. That’s a major step towards finding the killer app for LTE.AudiLTE_connect

However, those developments are in progress and have not actually resulted in any traffic yet, both are expected in 2014.

Consumers are driving traffic today. Looking through some public data, like Root Metrics reports on average data speeds, I am reminded by how inefficient the networks are.

Verizon vs AT&T LTE

Here are some results from gottabemobile.com that more or less concur the performance observed by Root Metrics.
iPhone 5 Verizon 4G LTE Speed Test Results
  • 8.52 Mbps download, 13.87 Mbps upload, 39 ms ping
  • 7.47 Mbps download, 13.70 Mbps upload, 38 ms ping
  • 10.21 Mbps download, 13.81 Mbps upload, 38 ms ping

iPhone 5 on AT&T 4G LTE Speed Test Results

  • 9.32 Mbps download, 13.69 Mbps upload, 47 ms ping
  • 9.11 Mbps download, 13.88 Mbps upload, 47 ms ping
  • 9.69 Mbps download, 16.20 Mbps upload, 48 ms ping
  • 8.96 Mbps download, 15.83 Mbps upload, 40 ms ping
iPhone 5 over Wi-Fi (Comcast Blast 50 Mbps) Speed Test Results
  • 8.52 Mbps download, 13.87 Mbps upload, 39 ms ping

So it’s hard for me not to notice a couple of elements to this story. Firstly, the mad rush to provide LTE coverage (schedule driven) has been, understandably, at the expense of quality. Here’s the difference:

LTESNR_chart

Important

The peak throughout of a 10MHz 2×2 MIMO with 3 PDCCH symbols, 5% extra overhead and 10% retransmissions channel is 58Mbps, with SIMO at 35MBps. So practical rates in the field are 26.4% of the peak rates with retransmissions. 

It doesn’t take a math genius to realize that 26.4% efficiency is an ROI buster.

So with 50,000 eNBs they currently carry maximally 750GBs over their RAN instead of the 2.9TBs they have already paid for. Simply extrapolating from Cisco’s 2013 VNI report, I get a projected data explosion like this:

CiscoVNI_extrap

Uhhh, isn’t the datapcolypse supposed to need about 800GBps  per operator in 2017? They’re blowing 2x that amount with poor efficiency!

 

 Making matters worse, ATT and VZW are starting to throw some additional spectrum into the markets. VZW recently announced they are adding 5000 eNBs with AWS spectrum, most likely ‘hotspots’ costing them $3.9B in a one time AWS spectrum buy plus ~$15K/eNB = $75M annually for the additional carrier. ATT has done a series of multibillion dollar deals to add in more 700MHz spectrum. 

ATT has been spending something on the order of $22B to build out and maintain an extensive WiFi network they claim offloads their LTE networks.

Here’s the rub, improving the efficiency to move their average is not as expensive as the offloading or spectrum additions they are doing. One more thing, SON should be doing some of the work for them in a very OPEX friendly manner. 

By cleaning up the network, making it more efficient, there is going to be an induced demand, yes, however having users use your product is not necessarily a bad thing. (I’ve already complained about pricing.)

Let me just say this is crazy. Reducing the network efficiency effectively increases your cost to provide service thus degrading margins. There are tons of levers to pull in the protocol as it is written, with improvements on the horizon. Wake up!

Ladies and gentlemen, the captain has illuminated the fasten seatbelts sign as some turbulence and or distruption is expected in this market.. 

PS: Verizon and ATT, I will make myself available to send you an invoice for, say, $1-2B and will turnkey a project to give you another 800GBps. You can sell your recent investments and return the money to working capital or the investors, or hold them for getting to 1GBps over the air with carrier aggregation in the future… (ok time for me to get back to work…)

Oscar

 

Tagged with →  
Share →
Buffer
VoLTE

VoLTE

Voice over LTE (VoLTE) is the next big thing. In fact, in 24 months, virtually all LTE enabled smart phones will support it. Curious?
Small Cells

Small Cells

Small Cells, previously known as 'femto' or 'pico' cells are possibly a savior to network operators. They offer capacity and coverage to the end user and are inexpensive for the network operator. Why aren't they everywhere?
Public Safety

Public Safety

LTE is and ideal technology for Public Safety use. See Why.