Results 1 to 11 of 11
  1. #1
    Join Date
    Nov 2003
    Location
    on the 'net
    Posts
    1,187

    Why is the Internet so screwy today?

    Something weird is going on with the net around the world today. Check out The Internet Traffic Report.

    53% packet loss in Asia right now and 666 ms average response time.

    25% packet loss in Australia.

    19% packet loss in North America, with about a third of the big routers showing in the red or completely dead.

    What are the possibilities:
    - a bunch of router upgrades gone bad?
    - hurricane messing things up?
    - cyber-terrorism?
    - sunspots?

  2. #2
    Join Date
    Jul 2003
    Location
    Earth
    Posts
    1,700
    Finally, South America is doing better than the other regions.

  3. #3
    Join Date
    Mar 2003
    Location
    New York City
    Posts
    7,406
    My guess would be hurricane, but I can be wrong.

  4. #4
    yea wow and if you look at the traffic it make a major drop out plus al ot of packet loss

    right now i am having major trouble connecting to my servers at ev1servers

  5. #5
    Join Date
    Nov 2003
    Location
    on the 'net
    Posts
    1,187
    It's pretty scary if one hurricane on one corner of one continent can make the net look like that!

    So much for building a network that can survive a nook-ya-lur war.

    I find it strange that the Internet Health Report is showing green across the board.

    I wonder if perhaps it's the InternetTrafficReport servers that are broken. Anyone know how that works?

  6. #6
    Hahahaha!

    This is easy.

    Lousy Internet Service Provider (Streamyx, in my country)!

  7. #7
    Join Date
    Jun 2004
    Location
    Northwest Colorado
    Posts
    4,636
    Not so easy, really. IHR reports network availability and latency, without taking into account dropped packets. ITR, on the other hand, is all about dropped packets. When packets are dropped, they must be resent. The resends kill actual throughput. This is why the Net is sluggish right now -- so much traffic is needing to be resent, that there is less bandwidth available for the next task.

    So while I've been down entirely for a short while today, mostly what I'm seeing is the Net behaving like all the links are saturated with traffic. It's just slow. But this most likely has nothing to do with your ISP or DC. Head on over to NANOG and start reading:

    http://www.merit.edu/mail.archives/nanog/msg12942.html
    http://www.merit.edu/mail.archives/nanog/msg13076.html
    http://www.merit.edu/mail.archives/nanog/msg13102.html
    http://www.merit.edu/mail.archives/nanog/msg13151.html
    http://www.merit.edu/mail.archives/nanog/msg13167.html

    I'm plowing through that info myself seeking enlightenment. Here's the most disturbing post so far:

    I wasn't thinking in terms of automatic monitoring, that would open up
    a can of worms security wise.
    Just looking at some way of getting the manual reporting (that is still
    taking place to the FCC) back in the (semi?)public domain. Due to
    terrorism concerns, that information is no longer available online. I'm
    pretty sure the LEC's and IXC's like it that way also, as they no longer
    have to air their dirty laundry. I was able to get some information
    under the Freedom of Information act for an outage that affected me
    directly , but that takes days or weeks. As close to real-time
    information as possible is what's needed to assess and respond to a
    major outage, i.e. routing voice/data via different carriers, being able
    to explain to end users why their email or phone calls didn't go through
    , etc. and eliminating the need to open tons of trouble tickets during a
    major event. One master ticket - such as fiber cut affect xxx OC48's
    would suffice.
    Not sure how this can be balanced against DHS perceived needs
    though...any suggestions?
    Make what you will of that, I do recall FCC outage reports used to be available online but it's been years since the last global brownout (if that's what we're even seeing). But I don't think the Internet being screwy has anything to do with Wilma, more likely those Level3/Verio routers Computerworld reported the other day:

    http://www.computerworld.com/managem...105678,00.html

    One company's borked routing tables sure can have a ripple effect, can't they? If anyone really knows what's going on please comment.
    Last edited by BigBison; 10-25-2005 at 01:41 AM.

  8. #8
    Join Date
    Jun 2004
    Location
    Northwest Colorado
    Posts
    4,636
    The insider scoop on Verio:

    Doug actually was seen buying some of that **** from an H-street john. Mother****er has NO damn shame. I'm out of here next week - just got hired, thank Christ. When the management is blowing doobies you know it's time to ****ing go.
    http://comments.****edcompany.com/phpcomments/index.php?newsid=107109&sid=&page=12&parentid=0&crapfilter=1

    Yeah, every backbone looks like it's dropping packets. But Verio still admits nothing in public. Threads like the above lead to questions of competence there to fix this problem. I'll tell you one thing I can barely access: my UltraDNS CP has been next to unusable for me for six days now. I've spoken to a tier 1 support rep there who was 'not at liberty' to discuss the Verio matter, but did confirm that my (and others') inability to use their CP wasn't their fault (other than choosing a suck-*** provider) nor is it mine, or my ISP's.

    Imagine what this is indirectly costing companies who, like UltraDNS, rely on Verio in terms of the surge in support tickets? I had to wait ten whole minutes to hear back from UltraDNS, with tier 1 support it's usually well under five. Apparently, they're busy, but again 'not at liberty' to discuss this. Huh. I will say that it's only my CP that's been problematic, my forward and recursive services are doing OK. But three of the four Verio routers I have to go through to access my CP are experiencing 20% packet loss at the moment, so how can my https CP session hope to work at even a moderately acceptable level?

    I still haven't seen any press coverage on this aside from the Computerworld article I linked to above. Anyone?

  9. #9
    Join Date
    Jun 2005
    Location
    The Server Rooms :)
    Posts
    163
    I haven't seen any news on it either.. I was talking to an admin at DedicatedNOW last night after getting some odd DNS issues with one of my servers (I thought it was DNS at the time) and he said they had some issues from slowdowns, etc..

  10. #10
    Join Date
    Jun 2004
    Location
    Northwest Colorado
    Posts
    4,636
    Hooray!

    ITR shows global packet loss has dropped below 1% in the past hour, over the past few hours the 'net has finally shaken itself off like a wet dog. I don't expect it's stable quite yet, but it's nice to see some throughput out here in the boonies for the first time since a week ago!

    Plus, I'm not dropping any packets accessing my UltraDNS CP right now, traversing the same route. This was not the case this morning. Nice job, potheads!
    Last edited by BigBison; 10-28-2005 at 05:44 PM.

  11. #11
    Join Date
    Nov 2003
    Location
    on the 'net
    Posts
    1,187
    Yes, it has all been pretty stable the last 24 hours or so.

    Any idea what happened, or why we are apparently the only ones who noticed?

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •