Were the Mayans right? Is 2012 the End of the Internet?

The latest episode of This Week in Startups was as always packed with interesting callers, two pitches (this time online education start ups Babbleflix and ThereNow) and a great guest.

James Segil, Co-Founder and President of EdgeCast a content delivery network gave great insides about the internet and on how video streaming could lead us to a breaking point.


You can watch the interview part above but I give you a short overview about the main points in this interview.

Without CDNs the internet would stop working
Content Delivery Networks are the duck tape that holds the internet together. They make copies of popular content and host it on POPs (Point of Presence) all over the globe so that content can be delivered in 80ms or less which is the time the human mind experiences as fluent loading. So this YouTube video above will be loaded from a server “near” you, depending on where you live.

From January 1st 2009 to January 1st 2010 the traffic has doubled
This is very impressive, because to get to the traffic number we had at the beginning of last year we needed about 20 years. Now we doubled this traffic in only one year mainly because of video streaming. This is Important for the next point.

The internet today has a reserve of 5x
We are still living on the infrastructure that was beefed up during the dotcom years. This worked well as the traffic grew constantly but relatively slow. But last year, as mentioned above, we doubled this traffic in only 12 months. At this rate we run out of headspace in about 2 years – 2012.

IPTV (TV over internet) did not even start
There are plans to stream TV, cable and satelite, over the internet in HD. This would mean huge amounts of traffic and the internet like it is today could not handle it.

Telecom companies are “under water” with heavy users
ISPs (Internet Service Provider) make no or very small money on heavy internet users. This of course means there is not much money left to invest into new infrastructure.

So if we count all of the above together, we might get into trouble. Obviousely we need to beef up the internet infrastructure in order to be able to handle the growing amount of traffic. This means there must be a ton of money invested in building new lines, server centers etc. But if the ISPs are already makeing not enough money today, what ever that means, to reinvest it into new infrastructure there is only one solution. The price of using the internet has to go up.

There are different scenarios possible. Number one of course, raising the price for internet access. This seems to be crazy for us as prices went constantly down all those years but charging the user is of course the easiest way.

Number two is charging the causer of the traffic, for example YouTube / Google. And in this case you can be sure that the costs will then be broken down to the users. For example your video is longer than 5 minutes and in HD, you need a premium account, only $9.99 per month. But wait a second, isn’t Google planning to build up their own highspeed network? Would their clients get better deals on video streaming? Google also posted a video with the title “Let’s make the internet faster” followed by a YouTube tool that shows you how slow your connection is.

And don’t forget the internet on mobile devices. There are rumors already that mobile operators are rethinking their prices because the usage of mobile streaming rose 99% between the first and the second half of 2009.

All of the above is of course also true for Amazon S3, Rackspace and all the other hosting companies. They will be charged by the ISPs and then they will break it down to their clients depending on the content they are delivering. And now we come to online education.

Virtual classrooms like Myngle, eduFire and Learn2Lingo stream audio and video. Language learning communities like Livemocha, Busuu and Babbel stream audio in their flashcards and exercises. New services like Sparkeo and Babbleflix stream huge videos. Hence they all might get higher bills of their hosting company which has to result in higher costs for the users.

This could become very critical for a lot of start ups because the calculation they did based on the idea that prices of webhosting remain steady or even go down further may turn into the opposite all of a sudden. And if they will be forced to charge more for using their services the result is quite clear.

Will this lead to the end of free or maybe even freemium on the internet? Of course it is easier for a company which already charges their customers from the beginning to higher their prices accordingly but for a company that offers big parts of the service for free changing to a paid only model might be tricky.

James Segil explains his thoughts on net neutrality and the idea that the internet needs to be free of charge in the video below.

The philosophy that the internet is free and that everything you get on the internet should be free is soon to come to an end.

It’s not possible for everything to be free. Answers shouldn’t be free, content shouldn’t be free.

So is this the end of the internet as we know it? What do you think? 

If you enjoyed this post, subscribe to the free EDUKWEST Newsletter and get updates via email.

With the subscription to the EDUKWEST Newsletter you will get
  • early access to upcoming interviews
  • news updates and reports
  • job listings in education 2.0
  • other exclusive perks

Your Email Address:

We respect your privacy and hate spam as much as you do. EDUKWEST will never sell, reveal or trade your email address.

  • Pingback: Kirsten Winkler

  • Chris Paul

    James Segil comments were quite interesting. I was surprised to hear that ISPs aren't doing as well as I'd thought. I think in order for ad-based sites to survive, they are going to have to create more effective advertising.

    Thanks for your roundup on Twist!

  • Pingback: Kirsten Winkler

  • Pingback: Mingoville

  • Pingback: Vikrama Dhiman

  • George Preston

    I'm glad that someone has the bravery to say that without doubt, the cost of internet access needs to rise. Otherwise we are paving the way for a return to walled-garden services where an ISPs will compete frantically for 'cheap and cheerful' access without multimedia provisions, or vendor-specific limitations on applications (e.g. “No Skype, but you can use AnyISPTalk instead” or even “No Youtube, why not try our online video portal instead”).

    Personally I think the content-can't-be-free argument and the problem of capacity are separate. The content industry and the infrastructure industry are of fundamentally different cultures, and in reality there is little overlap. They agree on almost nothing. We haven't seen film studios buying ISPs or ISPs buying record labels, as some predicted five years ago.

    Deloitte made some predictions along these lines back in 2007 (I can't find the actual original report but here is a reference article http://www.accountancyage.com/accountancyage/ne…) but backbones managed, although there was lots of hand-wringing and complaining to the BBC about iplayer.

    A perennial question in this debate is, what has happened to multicast? I work with delivering live video across enterprise networks for business television programmes. Without multicast, it's impossible. Switch on multicast, you can reach an enterprise base of 100,000+ without breaking [too much of] a sweat. True, it won't traverse an MPLS and is only feasable nationally not internationally…but doesn't iplayer fit perfectly with that? Multicast is *the* live layer of the model. Unicast is akin to 1000 cars all going from the same starting point to the same destination, en masse, to dispatch a letter. Multicast replaces them with one car and a photocopier at the destination.

    Forklift hardware replacement needed? Maybe. But we're already upgrading to IPv6! What next-gen backbone equipment is IPv6 capable but not multicast?

    Peer-to-peer (regardless of legitimacy) is reportedly consuming 60% of all bandwidth. But it is the most valuable part of the bandwidth – the last-hop upstream portion. The A in ADSL means that the uplink is a fraction of the downlink capacity, yet most peer to peer users' clients (think Spotify and 4oD before you think of bittorrent) do not respect this imbalance. There needs to be more intelligent peering which examines the topology so users in the same ISP network can crosstalk with each other without having to peer with other ISPs (that peering is a large cost to ISPs). Kontiki does this, but bittorrent certainly doesn't.

    Interesting times – it will be interesting to see how the model free-market web as it stands now will deal with the issues raised.

    George

  • BreslowAaron

    The end of the internet might have a more natural end than anyone thought. The year 2012 is said to have some powerful solar flares reserved for us. Along with solar flares that burts, a powerduf EM wave will hit the Earth. Every electronic device that does not have a shield against that… well, it's gonna fry… so bye bye internet…

  • liliag

    I discovered Livemocha a little over a year ago and although they keep the good stuff for paying visitors, they do offer very interesting and useful materials for those who need to learn a new language.
    Lilia Gephardt @ email hosting

  • Pingback: Kirsten Winkler

  • Pingback: dimi gineva