Log in

No account? Create an account
Previous Entry Share Next Entry
13 percent of *what*?
I'm going through today's haul of news articles trending on LinkedIn, and pondering this article on VentureBeat. The main point is the familiar one, that mobile devices are taking over the world. But it occurs to me that they may actually be underestimating the effect.

I was particularly struck by this quote:
Mobile devices now account for 13 percent of worldwide Internet traffic, up from 4 percent in 2010.
But what does "traffic" mean? Not all traffic is created equal.

I'm especially thinking about the fact that this article is, itself, absolutely *surrounded* by junk when I read it on my desktop. There are ads, links to other articles, a survey, social-networking connections -- all sorts of crap that is almost entirely uninteresting to me. I'd bet that that junk makes up 90+% of the bits on the page. Mobile sites, by comparison, tend to be leaner if they are well-designed -- on a small screen, you just *can't* have all that junk taking up space.

So here's the question. If mobile traffic is 13%, but is much more content-focused than desktop traffic, what percentage of actual *information* is going over mobile? It seems certain that it's more than 13%; I'd be fascinated to see a study that tries to suss out actual numbers...

  • 1
I expect that the bulk of that traffic is Audio and Video media, not static content found on web pages.

Hmm. Interesting point, and I suppose that's likely to be true...

The profile of overall Internet traffic has continuously changed over the years. In the packet forwarding business, I've kept close track of this, because changes in average packet size are a major driver in optimizing (and measuring) forwarding performance. Not to mention issues like sensitivity to latency, sensitivity to out-of-order delivery, and so on.

For these purposes, I'd primarily want to point out that measuring traffic in terms of bytes is very different than measuring it in terms of packets, and they're both important. Measuring the information content is a much more complicated issue yet, and my personal reaction to use Claude Shannon's definitions is little help in answering your question, which is implicitly defining information in terms of its utility to the end user.

And to look at the global numbers, umbran is completely right to suggest that readable text is a drop in the bucket anyway. Which doesn't mean that carriers don't try to give it special treatment, but compared to Overall Internet Traffic, it's lost in the noise.

Video accounts for lots of traffic

A large part of Internet traffic is video - I've seen estimates ranging from 80 to 90%. I'm sure many more hours of video are seen on PCs and smart TVs than on mobile devices. Video directed to larger screens is at much higher resolution and therefore takes a lot more bits than the very compressed video that goea to tiny mobile screens.

It would be interesting to know what percentage of non-video traffic goes to mobile devices.

  • 1