[outages] UUnet cross-USA latency increased

[Mark Kent]

Please note "FYI" at the beginning of my message.  I wasn't asking for help or analysis, I was just pointing out something that, 
based on our monitoring for the past nine months, is uncommon between the two direct-attached ports we monitor (in SJC and IAD).

Thanks,
-mark

On 10/18/12 10:02 AM, Jeremy Chadwick wrote:
> Mark,
>
> 1. Doesn't Verizon/MCI/UUnet/Alter.net use ICMP prioritisation on their
> routers?  Packet loss or latency seen at an individual router hop is
> usually (not always, but usually) indicative of that.
>
> 2. You stripped hops 4 onward from your output, which doesn't help
> analysis at all; if latency is truly a problem (oversaturated link,
> etc.), it "trickles down" through each subsequent hop,
>
> 3. You didn't provide a return-path traceroute (e.g. dst-->src) so it's
> hard to say "the problem is with Verizon" necessarily.  (I'm inclined to
> believe you given their history (believe me, I'm more than familiar with
> them at this point), but still...)
>
> This may benefit you:
>
> http://www.nanog.org/meetings/nanog47/presentations/Sunday/RAS_Traceroute_N47_Sun.pdf
>
> Finally, and maybe the most important point of all:
>
> If no loss is seen end-to-end (source-to-dest) then what's the concern?
> Or are you saying there's no packet loss end-to-end but there is latency
> end-to-end?  (Your Email doesn't explicitly say that, and I don't want
> to make assumptions...)
>