-- The ping spikes are still there. I have not explanati...

[mdk]

@LevonJay -- The ping spikes are still there.

I have no explanation why "96.52.32.1" appears twice in the output, but those ping-times seem to be relatively large, for packets that stay within your city -- namely a few kilometers from your cable-modem to the Shaw router in your city.

Shaw only has a few "cross-border" segments (Vancouver -> Seattle, Winnipeg -> Chicago). You can see this between the "CG" router and the first "WA" router, by the increase in the times.

Also, the segment leaving "SIX" to the next router definitely is "slow". There is nothing that Shaw can do about this.

[LevonJay]

I think using the cross-border segment that goes to Seattle rather than the one that goes to Chicago (where the servers are located) is the issue.

I think the first of the two 96.52.32.1 is a weird replacement of my local router which would normally show as 192.168.0.1? I'm not sure though.

It is definitely slow but it should also will not be hitting those segments if routed the other way. When it's working properly (which seems to be for one or two days every couple weeks, on Monday it was back to working fine then Tuesday changed again)

There does seem to be something wrong with even local pings, but I'm not sure if my area is still "off-spec"

I contacted Riot again and this is what they said:

I have some additional information that may be of use to the engineers with your ISP. I have seen a few different companies have issues with our gateways before. This is a link to a peering database : (https://www.peeringdb.com/net/5918). If your ISP has a peering agreement with us they will be able to contact our engineers and hopefully sort this out. My hope is that after providing them with that info they can get in resolve this. There is likely some configuration issue between our two networks that neither one of us can effect from where we are. I am sorry to have to bounce you between the two companies again, but there isn't anything more we can do on your system to help us out here.

[mdk]

@LevonJay -- but I'm not sure if my area is still "off-spec"

Again, that is a question that only Shaw Support can answer, by remotely logging-in to your cable-modem, to view the modem's measurements.

> on Monday it was back to working fine then Tuesday changed again

Who knows whether RIOT changed their routing-information, or if the routers detected a "fork in the road", and realized that going to Seattle is sometimes a "faster" route than going through Chicago, even though taking either "fork" will get you there.

[mdk]

@LevonJay -- at ‎2020-07-01 11:40 AM -- Why is not by default using the fastest route?

If you want to drive from Edmonton to Regina, you have at least two routes:

1. East on Yellowhead Highway through North Battleford and Saskatoon,
2. South through Red Deer to Calgary, then Highway 1 through Medicine Hat and Swift Current.

Which is the "faster" route?

Except in summer, it depends on the weather --  if you know that there is severe weather on one route, it would be reasonable to take the other route.

On Shaw's network, the amount of "network congestion" can vary from what is programmed into Shaw's router in Edmonton as being the "best" route to get out of Edmonton.

By using a VPN, you are forcing a third route.

Still, much of the latency is WITHIN Riot's network -- there is nothing that Shaw can do "inside" Riot's network.

[LevonJay]

It's been a while! Thought I'd give another update on the ongoing poor performance and issues.

Ping has been consistenly poor along with new bursts of random packet loss and larger than ever ping spikes. Local ping inside my city has stayed consistantly poor as well

The first charts are what me tracing the route to the game server.

For fun I thought I'd test what would happen if I pinged a hop that I knew existed on the optimal/correct route and see what that would look like, of course it looked like how the traffic should be routing with a much more stable and lower ping connection.

[mdk]

@LevonJay -- I thought I'd test what would happen if I pinged a hop that I knew existed on the optimal/correct route 

The Internet was designed as a "mesh" -- multiple routes to a destination, to survive one nuclear bomb.

So, I disagree with your "optimal/correct" classification.

Looking at both charts, the PING-times to "96.52.32.1" seem to be about the same, no matter which destination IP-address that you are tracing, but the values seem higher than expected, if you are physically located in Edmonton. Are you there, or in what prairie town are you? That "32.1" IP-address is a Shaw router in your town, which should be very quick to respond.

Also, in the second graph, the "hop" to that "ed.shawcable.net" seems slow, with a "current" value of "45.3". To me, this implies that you are not physically located in/near Edmonton. Correct?

So, you may have two issues, not just one:

1. slow PING values to that "32.1",

2. slow PING values to that "ed.shawcable.net" -- that segment between Edmonton and you could be quite busy.

[LevonJay]

It's definitively a more optimal/correct route to the game server: it is faster, more stable, with less hops. I'm not sure what other metric there could be.

I'm in northwest Edmonton.

That 45.3 is most likely a ping spike.

[mdk]

@LevonJay --  I'm not sure what other metric there could be

Each ISP announces its routing tables -- the path(s) necessary to reach their IP-addresses -- and each route has a "metric".

I've seen one case where packets from a Canadian University to CERN (Switzerland) were routed over the "Canarie" (fibre-optic) research network, but packets from CERN were returned over the "commercial" network, because the "metrics" preferred the slower ("commercial") Internet over the faster ("research") network.

> That 45.3

is most likely a ping spike

Show us the results from:  PING  -n  60  96.52.32.1

so that any "spikes" between your cable-modem and that Shaw router will be exposed.

Have you tried a different Ethernet cable between your computer and your cable-modem? A different coaxial-cable between your cable-modem and your wall-socket? Disconnecting the coaxial-cable, and reconnecting it, to ensure that it is tightly connected?

Have you tried the above PING command from a friend's Shaw Internet connection, to compare your results with theirs?

[LevonJay]

Shaw routing is incorrect. It's really that simple.

I pinged the 96.52.32.1 IP and this is the result. Which is pretty terrible. 

[mdk]

@LevonJay -- Shaw routing is incorrect. It's really that simple.

Shaw is CORRECTLY routing according to the routing information published by RIOT.

> terrible

I agree.

But, since this is a peer-to-peer discussion forum, not a direct path to Shaw Support, I recommend that you contact Shaw (currently, online chat has priority over telephone) and get them to trouble-shoot. They can remotely logon to your cable-modem, to view the signal-strength incoming to your cable-modem, and other values, and can check that ".1" router, too.

Have you tried a different Ethernet cable between your computer and your cable-modem?

Have you tried a different coaxial-cable between the wall-port and your cable-modem?

Are there other computers/devices in your home that are consuming your bandwidth?

Got any friends/neighbours in your area who also are Shaw customers? If so, get them to run the same test of that ".1" router, to view their results.

Do you have, or can you borrow, another computer, such that you can run your test on that computer, while connected to your cable-modem?

Can you take your WiFi-capable computer to a friend/neighbour's house, and try from there? Or, connect to the ShawOpen or ShawGo wireless networks, and run your tests?