said by hybrid02:Would you please post your traceroute google.ca and speedtest on your 25/10 and 50/10 package. Thanks
Addressing just the matter of traceroute (or ping) and [www.]google.c[a|om] in the context of Bell Canada:
1. Bell does very sketchy things with google, and with ICMP responses. I have been studying this for the past few weeks and am interested in others who care. (I realize that most don't.) For example, on two Fibe 10/25 networks, about 10% of the time (typically for about 300 seconds) ping -s 85 fails while ping -s 84 succeeds. Mutilated packets are returned! There is also Bell's failure to return proper ICMP echoes from (some of) its NS.
2. It makes a difference whether ping, traceroute, etc. are run from the modem, from the router, or from a LAN host, and also how the LAN host is connected to the router. (And, of course, other LAN traffic).
3. It should be understood that commands such as ping and traceroute are severely crippled in the Linux (Busybox) versions that run on most routers and modems. Tomato, for example. It is possible to get proper ping and other commands for a Tomato router.
(That said, mean latency to www.google.com is 4ms 90% of the time and 150000ms 10% of the time.)
IMHO this is related to the old Ikanos disaster and the new ping-no-browse manifestation.
I would be interested in comparing notes with others who have studied these things.