back to article Canadian ISP Shaw falls over with 'routing' sickness

Canadian ISP Shaw is experiencing a "nationwide outage affecting internet browsing", according to the polite phone robot put in place to discourage affected punters from overwhelming support systems. The issue appears to be multi-faceted. Your Canadian vultures have personally encountered numerous "rolling" outages over the …

  1. Peter Simpson 1

    One of our routers is broken...

    ...or we stuffed up a firmware update.

    But we can't tell you more, because we disabled remote access to all our routers as a security measure.

    // or some such tale

  2. Roo

    Have the NSA/Canadian authorities bungled their tapping operation ? :)

  3. NoneSuch Silver badge

    Have they tried turning it off and on again?

  4. Cipher


    ...has launched its invasion of Canada. Phase one: kill Command and Control...

  5. Anonymous Coward
    Anonymous Coward

    >> At various times Office 365 has been unavailable to both your northern vultures and various readers

    Office 364?

    1. BongoJoe

      It's fast becoming

      Office --364

  6. Dan Paul

    How dependent we have become!

    Fun stuff like this just shows how dependent we have become on technology.

    This is only the Shaw network, how is Rogers doing? Smells of poor router security to me.

    1. Trevor_Pott Gold badge

      Re: How dependent we have become!

      Rogers, Bell and Telus are fine...but anything that would transit the Shaw network due to routing or peering is pretty much blackholed. Which means the Canadian internet is pretty much borked.

      1. John Brown (no body) Silver badge

        Re: How dependent we have become!

        "Which means the Canadian internet is pretty much borked."

        What? You mean it's not routing around the problem?

        1. Trevor_Pott Gold badge

          Re: How dependent we have become!

          Nyet. There are so many fiddly little agreements, so many stupid little routes put in for political and financial reasons that the basis of routing on which the internet was founded - get the fucking packet there in the most efficient manner possible - seems to no longer apply. At least not for everyone, and certainly not all of the time.

  7. tempemeaty

    Ugh. Best of luck to all concerned

    Office 365, MS may control their own Asure Cloud's reliability...the connections in-between not so much. When I was doing tech support I can recall times when Office 365 outages cut us techs off from our hourly event notifications. The Shaw tech support guys must be getting hammered with calls and their own notifications may be cut off by this too. There's nothing like the experience of looking up at the call queue board and seeing 2000+ incoming calls waiting. Best of luck to Vulture North, Shaw and all their customers.

    1. Trevor_Pott Gold badge

      Re: Ugh. Best of luck to all concerned

      Aha, but what good is cloud computing if the network to gain you access is down?

      1. Anonymous Coward
        Anonymous Coward

        Re: Ugh. Best of luck to all concerned

        Network down? Surely you jest… networks don't go down!

  8. Barry Rueger

    Time for the daily outage I guess.

    Seriously, around here we just assume that once or twice a day the Internet will just stop happening for no obvious reason.

    Such is life with Shaw as your ISP.

    Personally I think that they're all pissed off at us for using Netflix at $9.95 a month instead of paying them $75 a month for a bundle of mostly crappy cable channels....

    (Is it just me, or has "routing issues" become the new catch-all excuse for any Internet outage?)

  9. Chris 3

    The 512 BGP Routes problem, presumably?

    My guess is that this is a symptom of the Internet reaching 512 BGP routes today. I'm not a router expert, but as I understand it 512 is the default max size on some older routers. Shaw isn't the only one having some problems today.

    1. Destroy All Monsters Silver badge

      Re: The 512 BGP Routes problem, presumably?

      Just like Y2K with less preparedness?

    2. Trevor_Pott Gold badge

      Re: The 512 BGP Routes problem, presumably?

      I'm investigating exactly that. The big question then is "why do we have some sites where all protocols work except a specific few?"

      I suppose it's possible that, for example, RDP (and not just to 3389, but all RDP!) is being sent to a DPI system and that hitting the 512K limit has screwed up routing for that protocol. I'll buy that as a possibility, but doing DPI on RDP sessions is really, really rude. I wonder if this didn't have some sort of cascade effect on DPI systems beyond just the basic routing issue.

    3. SixWordsLong

      Re: The 512 BGP Routes problem, presumably?

      Presumably so, except that it's 512 K (i.e. 512,000) BGP routes that was surpassed today.

      1. Chris 3

        Re: The 512 BGP Routes problem, presumably?

        Oops - quite so, thanks.

  10. Sarah Millin

    Shaw has been having weird problems for the last month. was blocked for several days. Any email domain they didn't like got blocked for a few days. I suspect they are trying out some new security features and breaking things along the way.

  11. Schrodinger


    There are widespread routing issues due to a lot of routers not being updated to allow for more than 512,000 entries in the BGP table.

    Short version: the internet is too big :-D

    1. Yes Me Silver badge

      Re: BGP

      Longer version: Filtering of long prefixes is going to get quite a bit more aggressive in the next few days/weeks/months. Some paths will get longer as a result, and some black holes will appear as a result.

    2. Anonymous Coward

      Re: BGP

      It's not that it's too big, it's that its too wide.

    3. localzuk Silver badge

      Re: BGP

      So, effectively, the BGP issue is caused by the lack of IPv4 addresses? As the larger blocks are reclaimed and chopped up into smaller subnets, and issued?

      1. -v(o.o)v-

        Re: BGP


        Several routers have FIB capacity of 512k routes as either hard limit or configurable limit where memory from other types (IPv6, VPNv4/6) can be reduced and given to IPv4. TCAM is expensive so lower end routers like 3B/3C non-XL models of Cisco 7600 PFC are at capacity.

  12. chris lively

    Sounds to me like they are doing DPI and playing with how to throttle certain traffic. They should probably call verizon and time warner to get some help

  13. json

    +1 on initial vulture analysis...

    ..if its a routing issue.. everything gets affected.

  14. razorfishsl

    So it's nothing to do with the bitcoin Hack then?

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like