Slashdot Mirror


User: Artefacto

Artefacto's activity in the archive.

Stories
0
Comments
28
First seen
Last seen
Profile
(view on slashdot.org)

Comments · 28

  1. FF, Opera and Safari have 3 modes too on Microsoft Confirms IE8 Has 3 Render Modes · · Score: 1

    Firefox, Safari and Opera have an "Almost Standards Compliant Mode" in addition to the Quirks Mode and Standards Compliant Mode. It shouldn't shock anyone that Microsoft is adding a third... See http://hsivonen.iki.fi/doctype/

  2. Re:Does filtering really work? on Interview with AT&T on BitTorrent Filtering · · Score: 1

    I'm pretty sure it is throttling.

    1) The global upload speed has high variability and, for long periods, averages way below my upload bandwidth.
    2) There's a strong correlation between average download speed and upload speed (more than was ti be expected)
    3) It started a couple of weeks after the limit of 30GB/month was removed.
    4) The behavior (severity of throttling) changes from time to time -- see http://web.ist.utl.pt/~ist155741/temp/04.PNG And no, I had no other bandwidth-hungry applications running. This happens all the time. Other protocols (e.g. HTTP) work fine and do not exhibit any of these patterns.

    Some people claim to have captured traffic between two ends and found extra RST packages, but, of course, I can't confirm that.

  3. Re:Does filtering really work? on Interview with AT&T on BitTorrent Filtering · · Score: 3, Interesting

    I'm not sure how they do it, but my ISP (netcabo, Portugal) is able to throttle bittorrent traffic even when the most strict encryption options are selected. They must be using good techniques to recognise bittorrent traffic in particular, because they're not able to throttle the encrypted protocols used by eMule. The method is so agressive that when it's in "throttle mode" other protocols may also be affected. It's also not mere passive throttling, they actually send false RST packages to your peers, so you can't keep a connection for long (talk about man-in-the-middle attacks...) And they don't acknowledge any of this ("p2p traffic speed is influenced by many factors" is the usual line) unless you get very deep in the support chain. There's no official position on this matter whatsoever.