Akamai Diversity

The Akamai Blog

Clarifying State of the Internet Report Metrics

In reviewing past media and industry blog coverage of Akamai's State of the Internet Report, we find that there has been some confusion about, or misinterpretation of, terms and/or data within the report.  In advance of the upcoming publication of the 4th Quarter, 2012 State of the Internet Report, we thought that this would be a good opportunity to provide some additional clarification.

As noted in a previous blog post, there are a number of sources of broadband speed data, including measurements collected by the United States Federal Communications Commission (Measuring Broadband America), Ookla (SpeedTest/NetIndex), Pando Networks, ChinaCache (ChinaCache Index), SamKnows, Netflix, and others.  As these, and Akamai's, measurements all make use of different methodologies, sample sizes, and user populations, they are all measuring slightly different things, though they may all use similar terminology.  To that end, there is not necessarily any one right or wrong set of data - each is right in its own way, considering the associated biases of the measurements.

The measurements published in the quarterly State of the Internet Report are taken from requests made to Akamai's HTTP/S platform.  (Over the course of a quarter, we're analyzing between 100-200 Trillion [yes, with a "T"] HTTP/S requests from over 200 countries/regions around the world.) To be completely clear, none of the data in the report comes from Akamai's native streaming networks (Adobe Flash, Apple QuickTime, Microsoft Windows Media).  As such, third-party assertions that the connection speed measurements published in the report are skewed lower by "rate-limited streaming" are incorrect - connections to our streaming servers are not part of the underlying data set used for the report, and Akamai does not throttle connections to our HTTP/S platform.  Adaptive bitrate streams are, however, delivered by Akamai's HTTP platform, but since they are delivered in a "chunked" fashion, they simply appear as a number of requests for content over HTTP. From a State of the Internet perspective, we're treating the delivery of video (adaptive bitrate streaming or progressive download) the same as any other piece of content we deliver via HTTP/S, like JavaScript, HTML pages, or software updates.

Within the report, data is presented at country and U.S. state levels - the geolocation of each unique IP address is done with Akamai's EdgeScape solution.  This data encompasses connections from both residential and business connections, and as such, represents a blend of the observed connection speeds.  (The report has never claimed to measure only residential traffic.)  Having said that, we believe that the number of residential connections to Akamai's platform greatly outnumbers the number of business connections, so the potential skew caused by higher speed business connections should be minimized.  In addition, as noted within the report, we endeavor to remove connections from known mobile networks from the underlying data set, so as to mitigate the potential skew caused by slower mobile connections.

Digging into each of the individual connectivity metrics featured within the report:

  • Average Connection Speed: As noted previously, this metric is calculated by taking an average of all of the connection speeds calculated during the quarter from the unique IP addresses determined to be in a specific country or U.S. state.  There are, however, a number of factors that can influence the average connection speed measurement:
    • Parallel Requests: According to the latest figures (4/1/2013) available from the HTTPArchive project, the average Web page requires 90 requests for content.  As such, a browser will likely open multiple connections to Akamai for various pieces of content, such as the base HTML page, a shared Javascript, or one or more advertisements.  Given that these requests are all competing for a limited resource (the Internet connection's bandwidth), each of these requests gets just a fraction of the overall resource, ultimately lowering the observed speed associated with each request.  (While these requests maybe sent over a single persistent connection from a browser to an Akamai server, they are each logged individually by Akamai.)
    • Small Files: Many of the components that make up a modern Web page (such as images, CSS files, Javascript files, AJAX responses) are relatively small in comparison to software downloads and media files.  As such, the connection used to download these smaller files is often short-lived enough that it doesn't exist long enough to reach maximum throughput rates.  (Think of it as being similar to driving on multiple short trips around your neighborhood on local roads vs. going on a multi-state roadtrip on the highway.)
    • IP Address Sharing: In this increasingly hyperconnected world, a single user may have multiple Internet-connected devices, and an average household today certainly does, from smartphones, tablets, televisions, and gaming systems to thermostats and refrigerators.  If multiple devices behind a single Internet connection (unique IP address) are all consuming content simultaneously, then as highlighted above, each device will ultimately have access to just a fraction of the whole connection, ultimately lowering the overall speed calculated for each request to Akamai.  (Obviously, this is also true on a much larger scale for business connections, where requests from hundreds of users may appear to be coming from a single gateway IP address.)
  • Average Peak Connection Speed: As noted previously, to calculate this metric, an average is taken of only the highest connection speed calculated from each unique IP address determined to be in a specific country or U.S. state.  Within the report, we note that we believe that the average peak connection speed is more representative of Internet connection capacity.  By using the fastest measurement observed from each unique IP address, we are capturing just those connections that reached maximum throughput rates.  Often, though not always, these connections are associated with the download of larger files, such as desktop applications, games, or software updates.  Furthermore, these connections are likely occurring late at night, or during some other period of lower usage, so while closer the theoretical maximum capacity of the connection, the measurements are unlikely to be representative of true throughput during normal conditions.  Furthermore, as an average, it will also mitigate the impact of the extremes - while there may be users connecting at Gigabit speeds in a given country, there will also be users connecting at dial-up, satellite, or DSL speeds.
  • High Broadband Connectivity: As noted within the report, this metric represents the percentage of connections to the Akamai platform, from a given country or U.S. state within a given quarter, which connected at speeds of 10 Mbps or more.  In order to "qualify" for inclusion in the rankings within the quarterly report, more than 25,000 unique IP addresses from a given country needed to make requests to Akamai at speeds above 10 Mbps in that quarter.  This threshold was instituted to prevent much smaller countries, with significantly fewer unique IP addresses making requests to Akamai, from artificially outranking larger countries with more Internet users.  Note that the High Broadband threshold speed was 5 Mbps from 2008-2011, and changed to 10 Mbps starting with the 1st Quarter, 2012 State of the Internet Report.
  • Broadband Connectivity: As noted within the report, this metric represents the percentage of connections to the Akamai platform, from a given country or U.S. state within a given quarter, which connected at speeds of 4 Mbps or more. In order to "qualify" for inclusion in the rankings within the quarterly report, more than 25,000 unique IP addresses from a given country needed to make requests to Akamai at speeds above 4 Mbps in that quarter.  Note that the Broadband threshold speed was 2 Mbps from 2008-2011, and changed to 4 Mbps starting with the 1st Quarter, 2012 State of the Internet Report.

Ultimately, the speeds an end user sees at any given point in time will likely be somewhere between the average and average peak connection speeds, but could be a bit faster or slower, depending on multiple factors.

In addition to the speed and broadband metrics, the "attack traffic" data published within the report also tends to be of significant interest, especially as "cyber" threats and Denial of Service attacks become increasingly frequent.  As noted within the State of the Internet Report, "Akamai maintains a distributed set of agents deployed across the Internet that monitor attack traffic.  Based on data collected by these agents, Akamai is able to identify the top countries from which attack traffic originates, as well as the top ports targeted by these attacks." These agents are unadvertised - that is, Akamai is not mapping any end user request traffic to them, nor should they be seeing any type of other traffic.  As such, we treat attempts to connect to the agents as attacks.  It is important to note, however, that the originating country as identified by the source IP address is not attribution of the attacker.  In other words, while Akamai may cite a specific country as being the source of a given percentage of attack traffic, the true attacker may actually be located elsewhere and using compromised systems in that country as a platform from which to launch the attacks.  As an example, a criminal in Russia may be launching attacks from compromised systems in the United States.  The attack traffic percentage published within the State of the Internet Report for a given country may serve as more of an indicator poor network/system hygiene (lots of compromised end-user systems) than as an indicator of how much the country serves as a haven for so-called "cyber criminals".  Note, too, that this attack traffic is separate and distinct from DNS- or Web-based attacks that may be targeting Akamai customer sites & applications - large attacks of those sorts won't generally be seen within this data set.

While Akamai does not publish data within the report at a network level, our network partners can check metrics for their Autonomous Systems (ASNs) on the Akamai Network Partner Portal.  If you're looking for data about a specific country that's not featured in either the global section or one of the regional sections of the report, check the Appendix, as it contains metrics for a larger set of countries.  In addition, we also suggest that you check the two data visualizations available at http://www.akamai.com/stateoftheinternet/ -- one displays metric values from the current quarter's report using a map-based interface, while the other enables users to graph historical values of key report metrics for the top 100 countries and all U.S. states.  Beyond that, as we note in each issue of the report, if you are interested in data that you can't find within the report, or if you have a question about something published within the report, please reach out to us via e-mail at stateoftheinternet@akamai.com or via Twitter to @akamai_soti

David Belson is the Product Line Director for Custom Analytics & MCDN at Akamai, and has been the primary author/editor of the State of the Internet report series since its inception.


Thanks for the clarifications.

About broadband connectivity: is it defined from average connection speed, or peak average connection speed?

In other words, could the rightmost column on page 45 say "% of IPs having above 4 Mbps as average connection speed"?