In the kingdom of the blind…

Posted in Regulatory, Spectrum at 12:50 pm by timfarrar

After discovering the retrospective revisions made to 2011 data traffic in Cisco’s Feb 2013 mobile VNI forecast yesterday, I wondered if this was something new, or if it was simply that no-one had ever noticed this issue before. Because some of the 2012 mobile VNI data is still accessible through the fixed VNI tool, it’s possible to do the same set of calculations for last year’s mobile VNI, and conclude that Cisco did in fact retrospectively change the 2010 data in Feb 2012 (on a global basis making an 8% increase for 2010 compared to the 13% decrease for 2011 seen in this week’s figures), as shown below:

The revisions to 2011 have the effect of enhancing Cisco’s announced data growth number for 2012 (now stated to be 70% on a global basis), because the estimated Dec 2012 traffic is only 48% higher than the 597PB/month for Dec 2011 traffic that Cisco estimated last year, and less than half the expected growth of 110% that Cisco forecast last year for 2012. Of course no-one is stepping up to admit that in fact Cisco’s new figures imply global traffic growth in 2011 was only 103%, rather than the 133% growth that was trumpeted originally as exceeding their own projections (and ironically this year’s press release doesn’t even mention that data growth was 70% in 2012).

It’s even more revealing to compare some of Cisco’s numbers to CTIA’s October 2012 statistics on mobile data traffic in the US. After all, even though CTIA has a blinkered focus on pushing the “spectrum crisis” (seen once again today in an op-ed quoting a statistic invented apparently out of thin air, that “Cisco forecasts mobile data traffic in the United States will be 20 times greater in 2015 than it is today”), at least their statistics are based on direct reporting by carriers accounting for 97% of wireless connections in the US, and so in this realm they ought to be king.

Given that wireless operators have no interest whatsoever in understating their traffic, its hard to see how US mobile data traffic could be higher than the CTIA statistics, yet that is precisely what Cisco’s estimates imply. Indeed Cisco have revised their Dec 2011 traffic estimate for the US upwards, to 127PB/mo, compared to the 108PB/mo estimate given in last year’s VNI forecast. However, CTIA’s statistics indicate that US mobile data traffic in the first 6 months of 2012 was 633PB, or an average of 105.5PB per month, and so if Cisco were accurate, then monthly data traffic would have had to decline between Dec 2011 and Jun 2012. Similarly, if Cisco’s Dec 2012 estimate of 207PB/mo of mobile data traffic in the US was accurate, then monthly traffic would have had to roughly double since Jun 2012, again an inconceivable shift (especially if traffic had supposedly fallen between Dec 2011 and Jun 2012).

Overall, based on CTIA’s statistics, I estimate that US mobile data traffic was roughly 96PB/mo in Dec 2011, and assuming consistent growth in the second half of 2012, the total for Dec 2012 should be around 134PB/mo, which is equal to a 40% traffic growth rate for monthly traffic during 2012. Of course that’s much lower than the 59% growth in total traffic between all of 2011 and all of 2012 that is implied by my extrapolation of the 2012H1 data, because the much smaller amount of traffic in the first half of 2011 pulls down the 2011 total. However, the 40% growth is the relevant figure to be used for comparison with Cisco’s numbers, and would have massive implications for the five year outlook: even a continuation of this 40% growth rate would represent only five-fold growth in monthly US mobile data traffic between Dec 2012 and Dec 2017, not the ten-fold growth that Cisco projects. The absolute gap would be even greater: roughly 720PB/mo in Dec 2017 vs 1963PB/mo forecast by Cisco.

Leave a Comment

You must be logged in to post a comment.