Archive for the 'Research' Category

Bandwidth Hogs Don’t Exist

Monday, December 5th, 2011

(or, A Decade into “Always-On” Internet.)

We’re in a paradoxical situation. broadband Internet networks keep getting better, but service provider limits on them keep getting worse.

As of this fall AT&T U-Verse has joined my cable Internet provider, Comcast, in imposing bandwidth caps of 250 GB per month on broadband Internet customers. Cox, Time Warner, Charter, and many more ISPs are now capping. Wireless or wired (and even on new fiber), the North American broadband landscape is now one where broadband internet means limits on usage.

The caps often come with additional fees. AT&T, for example, charges $10 per 50 GB as an overage charge in addition to the cap. Comcast simply threatens to permanently drop customers who go over the cap. As many commentators have pointed out, the ISPs seem to hate their customers, particularly when the customers like to use the product the ISPs are selling.

The alleged reason for the paradox of faster networks = more use restrictions is that the carriers claim the fat pipes have birthed a mythic beast, the “bandwidth hog.” A small number of “hogs” are taking up the whole Internet, and it just isn’t fair to you, say the carriers. The networks wouldn’t need such restrictions if only people would use the Internet “responsibly.” As the former satire news service Wallstrip put it, apparently the Internet isn’t big enough for everyone. Some of you have to leave.

We know from the research literature that there is a major psychological difference between metered and unmetered service. The initial rapid growth of Internet use in the US back in the days of the dial-up modem has been attributed to our historical preference for flat-fee local telephone service: calls to modem banks in the 1990s largely unmetered, and this left people free to play and experiment.

I was a research assistant for the authors of a key paper in the area eleven years ago.  The paper analyzed the consequences of an “always-on” unmetered broadband Internet–then a new idea. It pointed out that the ways people use unmetered always-on Internet are drastically different from those of metered use. Here’s the cite:

François Bar, Stephen Cohen, Peter Cowhey, Brad Delong, Michael Kleeman, John Zysman. (2000). Access and Innovation Policy for the Third-Generation Internet. Telecommunications Policy 24: 489-518. (Link to full text.)

It’s a decade later and things are going backwards. The new caps and metering are the opposite of our hopes for the Internet eleven years ago. Instead of “always-on” we’re now at “watch your meter” — if you can even find out how to check it.

Now, reporters say that a new study by consultancy Diffraction Analysis using actual ISP user data has found that bandwidth caps don’t work to improve congestion problems and that the mythical “bandwidth hogs” don’t actually exist. (See the excellent coverage at DSLReports.)

A quote from a reader of the DSLReports piece puts the facts in terms of the highway metaphor:

1% of vehicle drivers on the road travel a disproportionate amount of miles compared to the average driver. But they are on the road all the time. Most of the time they are on the road there is no rush hour congestion.The heavy drivers are likely to be involved in rush hour traffic jams, but only represent a small, not terribly relevant, fraction of total drivers in the traffic jam. Limiting the amount of miles a driver can drive does nothing to widen the roads and little to keep people off the roads during traffic jams, thus does not help with congestion.

In other words, given the weak correlation between a person’s bandwidth usage right now and a person’s total bits downloaded in a month, metering obviously targets the wrong thing and is not an effective tool for managing congestion.

To refine the highway metaphor a little bit, I would explain why bandwidth caps don’t work by saying:

Commuters cause traffic jams at 9 a.m. and 5 p.m. even if they only drive twice a day. Restricting the career travel of long-haul truckers won’t help.

Why do ISPs do it then? One answer is possibly incompetence. A variety of broadband carriers appear to struggle with the accurate tracking of consumer usage. They advocate metering but when they implement it they are unable to correctly measure how much data their customers use. They just don’t seem to have a grasp of what is going on inside their own networks.

However, surely many ISPs are aware of their own customer’s usage patterns. They know that “bandwidth hogs” don’t exist.  What, then, are they up to?

A much better motive for metering and caps in the case of triple-play operators (those who provide Internet and also provide pay television) is to protect their lucrative market for video from encroachment by Internet video competition.

In other words, Comcast has monopoly power in cable television in many markets. It has set its 250 GB/month cap so that it is impossible to buy television over its Internet service (take that, Netflix!) or to use free Internet video services in lieu of cable TV (take that, YouTube!), thus maintaining its monopoly in video. By the way, that is the definition of a violation of antitrust law. Hello, Justice Department?

The caps often don’t seem to have any connection to bandwidth anyway–at Comcast they were set at 250GB three years ago and have never been raised, despite drastic increases in Comcast service capacity with the adoption of newer DOCSIS 3.0 service and higher speed tiers. They also use the same caps across all of their service areas, which makes no sense if they are related to network capacity (which varies according to their adoption of DOCSIS 3.0, and other things).

It isn’t as though customers haven’t noticed this situation. Did you know that media companies–particularly carriers and networks–are almost all the most hated companies in national surveys of customer satisfaction? In one 2011 study, Internet service providers as a sector almost tied for last place (with newspapers), with Comcast dead last among its competitors at 59% satisfied.

It’s time to call out this farce of the “bandwidth hog” and the irresponsible user. Instead, let’s get some policies in place to control money hogs–the carriers that are exploiting monopoly power at our expense.


Updated 2:51 p.m. Fixed four typos, added the words “alleged,” and “fat pipes.” Added the commuters/long-haul truckers metaphor.

[This post was cross-posted to the social media collective. –CS]

Are Rural People Meaner?

Tuesday, September 20th, 2011

(or: Is Online Gossip a Question of Locale or Scale?)

I’m quoted in this morning’s New York Times Story, In Small Towns, Gossip Moves to the Web, and Turns Vicious. This came about because I’ve done some recent research on social media and rural communities (citations below), including a long-term ethnographic study of social media use in rural Native American communities in California and (with Eric Gilbert and Karrie Karahalios) a study of rural vs. urban use of social networking sites.

Here’s the story in a nutshell: In case you aren’t aware of the web aggregator (, it is a portal site owned by newspaper companies that provides a “home page” for every city and town in the US.  That page consists of a feed of local news, presumably generated algorithmically, mixed in with weather, polls, and–critical for our story today–a forum.

Topix forums have become the online place to be for some small towns.  Unlike social media sites such as Google Plus and Facebook, which have pursued a policy of only allowing real names online, the Topix forums allow anonymous posting.

The result is a cesspool of gossip, with posts that have titles like, “People to Stay Away From.” That thread consists simply of a list of the real names of people in Pearisburg, VA that the poster, a “Mr. Kickass,” doesn’t like. The NYT piece included some examples but they chose tame ones.

A normal Topix small-town board includes a purported attempt to out a gay man, accusations that so-and-so has AIDS, a diatribe against miscegenation, public shaming of “bad parents,” announcement that this or that person is a crackhead, and more.  All of these posts are anonymous.

Yes, it looks like Topix is transforming gossip. Just as the Internet has transformed buying airline tickets, is making gossip more efficient. It’s now easier to reach a larger number of people, and (a terrible side effect), indexing by search engines means that an impermanent medium like gossip can now stay online indefinitely to haunt you forever. And I agree that gossip can ruin lives. There are problems.

Yet it’s not clear to me that these are rural problems. I agree that rural people are different from urban people. They are in aggregate more likely to be older, less mobile, poorer, and less educated. And we know that rural people use the Internet differently from urban people.

But remember that Juicy Campus scandal about three years ago (NYT: College Gossip Leaves the Bathroom Wall and Goes Online)? This was a new online forum that allowed anonymous posting, and it filled up with scandalous gossip about sex and drugs (well, mostly sex). It ruined lives. That was a scenario but the locale wasn’t small-town America, it was the University of California, Duke, and Yale.

Chris Tolles, the Topix CEO, is quoted in the Times article linking the situation on the Topix forums to the Hatfields and McCoys. C’mon, Mr. Tolles. Give us a break. At least he didn’t mention Deliverance.

I think the formula is:

anonymity + a defined community (scale) = gossip

Rural doesn’t appear in that equation.

Champaign-Urbana, where I live, is a small town, but it is too big to fit most definitions of a rural area. I think it would be great if all of the gossips, racists and bigots lived on farms somewhere far away from me, but I just don’t think that’s the case. (For more on this, see Mary Gray’s excellent book.)

The situation as a whole reminds me of early efforts to spread the telephone to rural America a century ago (see Fischer’s excellent research). Then, CEOs of telephone companies often refused to build in rural areas because they thought that rural people were all poor and stupid. All of the major telephone company CEOs lived in big cities, and they were sure that rural folks, if given a telephone, would be too dumb to use it, would complain a lot about it, and would probably only play banjo to each other anyway.

I don’t think rural people are meaner.

[This post was cross-posted to the social media collective. -CS]

[Thanks to Kristen Guth for thinking of the Juicy Campus comparison.]


Further reading:

Sandvig, C. (2012). Connection at Ewiiaapaayp Mountain: Indigenous Internet Infrastructure. In: L. Nakamura & P. Chow-White (eds.) Race After the Internet. New York: Routledge. (link to proofs)

Gilbert, E., Karahalios, K. & Sandvig, C. (2010). The Network in the Garden: Designing Social Media for Rural LifeAmerican Behavioral Scientist, 53 (9): 1367-1388.


There’s More to Internet Freedom

Sunday, March 6th, 2011

(or, Is Google the new United Press Syndicate?)

Over at the Huffington Post I have a new blog post co-authored with Dan Schiller that tries to flag the current and timely Internet Freedom / Free Flow of Information / Right to Connect debate with a little more context. Here it is:

Free Flow of Information and Profit
by Dan Schiller and Christian Sandvig

(Photo tweeted from Cairo’s Tahrir Square by @richardengelnbc
click to enlarge)

In at least some sense it is a response to Monroe Price’s earlier post, “Clinton’s ‘Long Game’ Advancing Internet Freedom,” as well as Ethan Zuckerman’s reaction to Clinton’s 2011 Internet Freedom speech.

HuffPo made it a “Featured Post!”  Woo!

But there are no comments.  Boo! I think we used too many big words.

This photo is from Richard Engel of NBC.  Translation is supposedly “thank you / facebook / youth of Egypt.”

Internet Innovation: The Big Read

Monday, August 30th, 2010

(or: An Internet and Innovation Reading List for You.)

Recently I’ve been ginning up a reading list about the title given in this blog post, and I wondered if I could try to crowdsource some of this bad boy. If you had a semester and you wanted a graduate-level someone to learn all of the basics and some of the more advanced and interesting stuff about the broad topic “the Internet and innovation,” what would you tell them to read?

[The preview image for the
“History of the Internet”
video by Melih Bigil.]

Read the rest of this entry »

Technology Studies Needs Both Priests and Missionaries

Friday, July 2nd, 2010

(or: Your technology may be political, but who cares?)

I am a longtime admirer and participant in the intellectual crossroads known as Science and Technology Studies (STS). I first read Langdon Winner‘s “Do Artifacts Have Politics?” and he changed my life.  In my first job I had the chance to organize a conference around my own research interests, and my choice for the keynote was Steve Woolgar (“Laboratory Life,” “The Machine at Work,” “Science, The Very Idea,” “Virtual Society?”).

My intellectual links to STS were formed in  graduate school but I continue my involvement in an STS program now as a faculty member.  I really enjoyed the last 4S conference, where I presented some of my recent work, and I’d like to think it was well received.  I say all this to introduce this post because I want to emphasize that I feel like an STS insider.

In the last few weeks I’ve had a conversation like the one I’ll describe below with three other STS insiders, so I thought I’d share it.  Tell me if I’m crazy.

[Click to enlarge — original photo by lawprier on flickr]

If you’ve read any STS, you know by now that STS scholars typically make arguments of the form: “What you thought was technical is actually political!” Read the rest of this entry »

Bad Behavior has blocked 509 access attempts in the last 7 days.