Slashdot Mirror
Think-Tank Warns of Internet "Brownouts" Starting Next Year
JacobSteelsmith writes "A respected American think-tank, Nemertes Research, reports the Web has reached a critical point. For many reasons, Internet usage continues to rise (imagine that), and bandwidth usage is increasing due to traffic heavy sites such as YouTube. The article goes on to describe the perils Internet users will face including 'brownouts that will freeze their computers as capacity runs out in cyberspace,' and constant network 'traffic jams,' similar to 'how home computers slow down when the kids get back from school and start playing games.' ... 'Monthly traffic across the internet is running at about eight exabytes. A recent study by the University of Minnesota estimated that traffic was growing by at least 60 per cent a year, although that did not take into account plans for greater internet access in China and India. ... While the net itself will ultimately survive, Ritter said that waves of disruption would begin to emerge next year, when computers would jitter and freeze. This would be followed by brownouts — a combination of temporary freezing and computers being reduced to a slow speed.'"
I remember this from an earlier slashdot of the same group saying the same thing. http://slashdot.org/article.pl?sid=07/11/20/0024248&from=rss
Respect in this case comes from the Internet Innovation Alliance who fund it. Of course, AT&T funds the IIA
Make of that what you will. I know that the first thing I think is "shill", followed closely by "astroturf".
Watch for this study to be cited in some bills regarding tiered service agreements any day now.
No, of course it wouldn't - not unless your web browser is poorly written and stuck in an I/O blocking state, consuming all available CPU cycles. But that doesn't happen these days, and hasn't for a decade+. Never mind the bravado in which the article states these things is, and always has been, nonsense.
~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
Take a look at why Slashdot's pages load so slowly. There are several layers of "document.write(some javascript that loads something else)" just to load ads. The browser can't do the loads concurrently; they all take place sequentially. Each "document.write" has to finish before the code in it can be run. Also, some of the CSS is being read from "s.fsdn.com", which is a rather slow server at times.
It can get worse. Try Rushmore Drive, the slowest-loading search engine home page known. This is a spinoff of Ask. There's enough ad-related crap on that page that it takes 10-15 seconds to load. And this is without any personalization or content-related overhead. It's all inept ad serving.
Those are both sites maintained by supposedly competent professionals. Sites where some third-tier web programmer just cut and pasted code from other sites can be much worse.
We can probably deal with increases in Internet traffic just by improving ad-blocking.
The summary was so bad that I actually read the article, expecting that I could then come here and post the usual flame about mangled, misleading, or otherwise just bad summaries.
That was a HUGE mistake. The article really is bad enough that no improvement in the summary would have been possible.
The author of that article confuses "computer" and "network streaming". The confusion seems to be quite deep, perhaps to the point that the author thinks of computers as mere display screens for this magical "internet" thing that does all the work.
Imagine that you read an article about a traffic jam, but rather than saying that the flow of traffic at the moment didn't seem to be very fast, it instead suggested that the cars would "jitter and freeze". That's how I felt when I read that article.
See that "Preview" button?
For example, 1000 connections each going at 10k a second (not unreasonable numbers) = about 10,000k of transfer trying to come into the ISP. It doesn't matter if they're filtering it down to 128k/sec or whatever you're paying for -- that's still 80 megabit worth of bandwidth resources wasted on the ISP's side.
WTF? TCP doesn't work like that... The sending speed changes according to the acks the receiving end sends back. The ISP gets exactly what it sends to the user and everybodies happy.
You're right though that the overhead of a couple thousand connections can be quite large - but thats not the problem here. The problem is that the same slice of bandwidth is sold to 10 different people. This just will not work.