Cisco Blogs


Cisco Blog > SP360: Service Provider

Announcing the 2010 Cisco VNI Usage Study Findings

Few things get me as fired up about this industry as being able to dive into some serious data about, well, data. By now, most of our readership is quite familiar with our Cisco Visual Networking Index study or VNI. It’s comprised of several separate efforts, with the VNI Forecast being the best known of the bunch and the one that is most often quoted in the media, used by regulators, and factored into architectural plans by our customers. The Forecast is, as its name conveys, offers a forward looking view. Using third party subscription growth forecasts and advanced modeling techniques, we deliver an estimate of the amount of traffic that will be crossing global IP networks over the next half decade.

But what is real traffic looking like now, you ask?

For that, we have the Cisco VNI Usage study - unlike the Forecast, which uses third-party subscription growth inputs, VNI Usage assess the activity taking place over networks every day.  It’s a compilation of anonymous data sourced from millions of broadband subscriber lines from over 20 service providers worldwide.  While we don’t know the names of the user or the specific content viewed, this unique, primary data does provide subscriber-level insights into a variety of factors such as: the type of applications used; when and how much bandwidth is consumed; and a context for the networking trends and challenges that many network operators are grappling with today.  It’s real data on data that’s real interesting (and you can quote me on that).  Over a series of several posts this week, I’ll dive into details of the Cisco VNI Usage study that was just released this week and covers the research period of the third calendar quarter of 2010.

Here are some top takeaways:

  • The average broadband connection generates 14.9 GB of Internet traffic per month, up from 11.4 GB per month last year when we ran this same study - this is an increase of 31 percent when averaged out across the global subscriber base.
  • Peer-to-peer (P2P) file sharing is now 25 percent of global broadband traffic - last year it was 38 percent of total traffic.  It’s important to know that despite this significant drop in percentage, the overall about to traffic generated by P2P in absolute terms is still growing - it’s just growing more slowly than visual networking and other advanced applications such as online video.  Which leads me to…

  • Peer-to-peer has been surpassed by online video as the largest category. The subset of video that includes streaming video, flash, and Internet TV represents 26 percent of total traffic generated, compared to 25 percent for P2P.   The VNI Usage 2010 whitepaper provides granular breakdown of the broadband traffic generated by application category and subcategory, from specific gaming consoles to voice calls.
  • The top 1 percent of broadband connections is responsible for more than 20 percent of total Internet traffic. Amazing the impact on the network that a group of “uber users” can have…
  • Expanding this notion further, the top 10 percent of connections is responsible for over 60 percent of broadband Internet traffic, worldwide. These insights provide tangible context around the issue of tiered pricing plans.  Should the “all-you-can-consume” broadband pricing model continue or are the 90% effectively subsidizing the usage of the 10%.  What do you think?  And of note, despite the increase in overall broadband consumption from last year to this year, this percentage consumption of the top few percent remained the same.
  • In an average day, Internet “prime time” ranges from approximately 9 p.m. to 1 a.m. around the world. This contrasts to broadcast TV prime time, which is generally from 7 p.m. to 11 p.m. across most global markets.

This is an interesting one, especially when we look at hour-by-hour consumption trends. I’ll cover that tomorrow, but in the interim, please check out the findings more yourself and get statistics that will help ground theoretical discussions around networking trends with some real-life data.

Until tomorrow!

Tags: , , , , ,

In an effort to keep conversations fresh, Cisco Blogs closes comments after 60 days. Please visit the Cisco Blogs hub page for the latest content.

3 Comments.


  1. I can’t wait to see the actual data. If the data gets into BGP and peering, then it’ll be a home run.

    Without this feedback loop, Cisco’s VNI was starting to look like a well run marketing program driven by a set of yet-another-project-and-fill-right analysis tools. The source data being IDC’s market numbers and comScore’s media measurement.

       0 likes

  2. Doug Webster

    Hi Daniel,
    Glad you found the study worthwhile and I encourage you to check out the whitepaper on it. We won’t be publishing the raw data so I don’t know that you’ll have you BGP need fulfilled… but do hope the rest of the data is beneficial to you, too!
    We’ve been using this as an input means for the forecast for sometime now and this is the second year we’ve published the VNI usage, so I would argue your point that the source data for VNI Forecast was two sources from the onset since it entails a good many more and ends up being quite a complex modeling effort. If only it were so easy! We detail the methodology of that in a separate whitepaper on the site and consistently look for further enhancements when we discuss the findings with regulators and customers worldwide — we’d certainly appreciate any suggestion you have as well for future versions of the study.
    Lastly, I had the chance to read your blog and glad you are positive on the effort. On your point re. “Internet Prime Time” though, please don’t look at it as a measurement of the number of user viewing any or all content on the web as input for an advertising perspective. Rather, we view it as a measure of traffic volatility where there can be significant spikes in the amount of bandwidth consumed. This data is very much of interest to network architects as they are not able to plan on their network to just handle the growth mapping to the Forecast findings in a linear manner but also need to factor in the spikes in traffic that will occur in order for their customers to get the desired level of experience.
    Thanks again for your comments and readership,
    Doug

       0 likes

  3. Doug — Fair enough.

    It’s only a matter of time before traffic engineering will be inseparable from media measurement. I guess it’s the mixed blessing of an all-IP world.

    Daniel

       0 likes