avatar image

Welcome, Guest. Please login or register.
Did you miss your activation email?

Login with username, password and session length
  • January 16, 2019, 10:49 AM
  • Proudly celebrating 13 years online.
  • Donate now to become a lifetime supporting member of the site and get a non-expiring license key for all of our programs.
  • donate

Author Topic: Traffic Growth Could Choke 'Net by 2010  (Read 2119 times)


  • That scary taskbar girl
  • Global Moderator
  • Joined in 2006
  • *****
  • Posts: 5,765
    • View Profile
    • Donate to Member
Traffic Growth Could Choke 'Net by 2010
« on: November 20, 2007, 05:12 AM »
One of the great things about technology is that it's surprisingly easy to underestimate how strong the demand for it can become. Remember IBM CEO Thomas Watson's purported 1943 remark, "There's a world market for maybe 5 computers," and Bill Gates' supposed 1980 comment, "640 kbytes of RAM ought to be enough for anyone"?

Interestingly, there's no evidence that either is an actual quotation -- and Gates in particular has emphatically denied the words attributed to him. However, they linger in the public consciousness as a way of reminding ourselves of a greater truth: The appetite for technology can surprise even seasoned veterans.

That certainly happened to me recently. Awhile back, my colleagues and I decided to look at how demand for Internet capacity might change over the next five years, and see whether the existing and planned infrastructure is adequate. To do so, we modeled user consumption of bandwidth over time, validated it against the best available data and then projected that demand forward. It's essentially a Moore's Law model of Internet demand, in that it looks at the rate of increase in a commodity (processing power in the case of Moore's Law, and bandwidth use here). Then we compared that demand to existing and planned infrastructure capacity...

Ralf Maximus

  • Supporting Member
  • Joined in 2007
  • **
  • Posts: 927
    • View Profile
    • Read more about this member.
    • Donate to Member
Re: Traffic Growth Could Choke 'Net by 2010
« Reply #1 on: November 20, 2007, 08:33 AM »
My favorite quote from the report itself (full text here):

One could even whimsically speculate—as we did in the title--that the lack of investment could be holding back the time at which the Internet reaches a “singularity” (a point at which accelerating change creates an unpredictable outcome, such as the Internet becoming independently sentient).

So throttling the internet may be a good outcome, if we don't want a real-life Terminator movie.  Unfortunately, things go downhill after that...

Basically, they're saying that at current consumer consumption growth rates, combined with a continued growth in the appearance of new web 2.0 apps, the internet will choke on the traffic within 3 to 5 years.  Mark.

Note my emphasis of "growth" up there... that's key to their whole study, which they acknowledge early on.  However, that particular metric bugs me for a couple of reasons.  In section 4, "Modelling User Demand" they describe their methodology for predicting how much bandwidth a typical internet user will consume in the future.  They base this on a dataset of consumer devices purchased from 2000-2007, add up all the bandwidth each device can use, then throttle it back by the max speed of consumer internet gateway components.  Their findings:

We found that, counting all devices and all port interfaces, an Internet-attached North American user in 2007 is theoretically capable of generating approximately 61 Mbit/s of traffic, equating to 20 Petabytes of traffic per month.

*TILT*  Uh... what?  My Comcast "business internet account" has a max down speed of roughly 6 Mb/s, or 10x less than their number.  Things are already sounding fishy.  But I do like the 20 Petabytes of traffic a month; that sounds sexy.

But wait, they acknowledge these are top-end numbers, and adjust accordingly:

It’s important to note that by this measure, not only does maximum possible demand exceed Internet capacity today, but it has always exceeded capacity and very likely always will, because of the fact that port speeds on Internet-attached devices tend to be within an order of magnitude or so of Internet circuits themselves (and there are far more devices than circuits).

Wait, no!  That's even worse!  Where's the "yes we know using this 'theoretical max' number in actual calculations is stupid" statement?  Ah, here we go...

Moreover, access circuits (the last mile typically connecting Internet-attached devices to the Internet) tend to run at speeds several orders of magnitude lower than the devices they serve. A typical DSL circuit, for example, delivers a maximum of 1.5 Mbit/s to one or more Internet-attached devices in each household (and each device, keep in mind, has a port speed of 100 Mbit/s or up). This means that the Internet-attached devices in a typical household are fully capable of saturating the household’s Internet connection; the same holds true for business sites.

Ahhhh!  Go back!  You're making it worse!

However, simply because the devices can saturate the “Net doesn’t mean they will. In practice, network-attached devices very rarely (in fact, almost never) generate traffic at 100% of port capacity for a sustained period of time.

Thank god.  They've come to their senses.  Well, not really... because that's all they say on the subecjt before skipping onto the next section, "Predicting How Much Porn Users Will Download".  Well, it's not really it... more like an analysis of how users will use the choked, Terminator-controlled internet of 2010.  There's a discussion of web 2.0 impact and YouTube's new "We Give Up" initiative where they start streaming pornography in 2009, which crashes all their servers and frees up an additional 25% bandwidth which is one reason why things still work in 2010.

But I digress.

The authors really seem hung up on the "port speed" of the device (100 Mb/s on my PC), not the at-the-curb demand made as measured on the public side of the gateway (6 Mb/s here) which, when you're predicting the doom of the internet is the only number that really matters.

It's a bit like freaking out because all those people riding a bus might go buy cars and overrun the freeway system.

So to summarize, the internet will choke on consumer traffic sometime in 2010, but only if every projected household upgrades their broadband speed by a factor of 10, then saturates their connection 100% continuously 24/7 between now and then.

Aieeee!  Run for the hills. 


  • Charter Honorary Member
  • Joined in 2005
  • ***
  • Posts: 9,134
  • [Well, THAT escalated quickly!]
    • View Profile
    • f0dder's place
    • Read more about this member.
    • Donate to Member
Re: Traffic Growth Could Choke 'Net by 2010
« Reply #2 on: November 20, 2007, 09:06 AM »
Humm, traffic grows and it does so at a pretty fast rate, but I don't really see a problem - capacity increases as well. And I bet the two largest quantities of traffic would be spam/DDoS and warez.

For warez, DVD-R movies weigh in at ~4gigs a whop, games are between 4-10gigs, and then there's those insane hd-dvd/bluray releases at 20gig. It's not uncommon to see home users with 10mbit or faster lines (20/1mbit adsl is quite affordable in .dk, for instance). If you sum up all the warez activity on the net, I wouldn't be surprised if it was in the multi-gigabytes per second range. Constantly. Today even the public p2p torrent tracker communities have several "seedboxes" with 10/10 or even 100/100 mbit lines. Back in early 2000, 100/100 private servers wasn't uncommon for the so-called "scene" oh-so-private ftp servers, and today I guess the top-topsites are on gigabit links.

And yet that doesn't affect my access to or other sites - only thing that tends to limit my web browsing experience would be a site on a shared host with an overloaded database server.

People are painting the devil on the wall, yelling that on-demand HDTV broadcasting will kill net bandwidth etc... but it won't. You'll receive your content from your ISP on your fiber line (remember that each individual fiber strain can do at least 1gbit/s full-duplex, and they bunch a lot of those together), and that traffic doesn't need to go through the internet backbone and take up traffic, it goes through your ISPs internal way-high capacity backbone. And do remember that all the warez traffic, generating insane amounts of traffic, is routed on the internet, without without too much trouble.
- carpe noctem