nanog mailing list archives
Re: Spitballing IoT Security
From: "tim () pelican org" <tim () pelican org>
Date: Thu, 27 Oct 2016 09:53:31 +0100 (BST)
On Thursday, 27 October, 2016 00:40, "Ronald F. Guilmette" <rfg () tristatelogic com> said:
Point: I have a DSL line which is limited to 6Mbps down and 756Kbps up. My guess is that if any typical/average user is seen to be using more than, say, 1/10 of that amount of "up" bandwidth in any one given 10 minute time period, then something is really really REALLY wrong.
This sounds like a horrible view of the Internet as "TV, only with more funny cat pictures", where most users are in a second-tier that is only expected / allowed to consume. One of the reasons I'm very grateful for FTTC / VDSL is that I can finally get a useful upstream speed. Going from 10-14M downstream to 80M was very nice, but going from 1M to 20M upstream was an absolute game-changer. I back up to the cloud - and there are plenty of services that allow regular, non-technical users to do this. The initial run saturated my upstream for days, and the incrementals are sometimes 20 or 30 minute bursts. I wouldn't even have tried on ADSL. Every time I get back from a day out, or even more so from a holiday, I upload the photos from my PC to one or more cloud services. I'll max my uplink for anywhere between 10 minutes and an hour - on the old ADSL, it was easily an overnight task. Working from home, I can now work directly with files on network shares, rather than copying everything to the laptop before I leave the office and trying to sync changes when I get back. I know the exact figures for this case, but there are a *lot* of spikes over the course of a day. With ADSL, I could go and make tea every time I needed to save a large Word doc or Powerpoint back to the network. On top of that, I can spend anything up to 3 or 4 hours in videoconferences, which will have a steady stream of a few hundred Kb/s. Spotting atypical (or ideally malicious) traffic is a valid goal, but I think we need to be a whole lot smarter than "customer is using upstream". Regards, Tim.
Current thread:
- Re: Spitballing IoT Security, (continued)
- Re: Spitballing IoT Security Jon Lewis (Oct 27)
- RE: Spitballing IoT Security Emille Blanc (Oct 27)
- Re: Spitballing IoT Security Rich Kulawiec (Oct 28)
- Re: Spitballing IoT Security Geoffrey Keating (Oct 27)
- Re: Spitballing IoT Security John Levine (Oct 27)
- Re: Spitballing IoT Security Leo Bicknell (Oct 27)
- Re: Spitballing IoT Security Ronald F. Guilmette (Oct 26)
- Re: Spitballing IoT Security Chris Boyd (Oct 26)
- Re: Spitballing IoT Security Mark Andrews (Oct 26)
- Re: Spitballing IoT Security Mel Beckman (Oct 26)
- Re: Spitballing IoT Security tim () pelican org (Oct 27)
- Re: Spitballing IoT Security Ronald F. Guilmette (Oct 27)
- Re: Spitballing IoT Security knack via NANOG (Oct 27)
- Re: Spitballing IoT Security Leo Bicknell (Oct 27)
- Re: Spitballing IoT Security Ronald F. Guilmette (Oct 27)
- Re: Spitballing IoT Security Ken Matlock (Oct 27)
- Re: Spitballing IoT Security Ronald F. Guilmette (Oct 27)
- Re: Spitballing IoT Security Laszlo Hanyecz (Oct 27)
- Re: Spitballing IoT Security bzs (Oct 26)
- Re: Spitballing IoT Security Valdis . Kletnieks (Oct 26)
- Re: Spitballing IoT Security Josh Reynolds (Oct 26)