Need help with connection
|
it looks to be a peering issue between level3 and network layer in Miami.
|
|
" Ohai Drakier. Yeah, it seems things get funky between them every now and then. I don't think we can do much (ISP or POE guys) except waiting for this 2 companies to fix their peering issues :( EDIT: Edited previous post with another MTR of another server. Last edited by daemmon#1312 on Nov 4, 2014, 6:22:20 PM
|
|
|
You are correct. These types of issues are harder to escalate and can really rely on your ISP for that (which many won't do).
Sorry. |
|
|
its horrible right now, im having 500ms.
There is no real way to fix this? Who decides the route by USA? my ISP or some company from the USA? |
|
" It's a network of peers. The peers determine the routing. You connect to your ISP. Your ISP connects to a bunch of peers (depending on where you're going). Those peers (carriers) also connect to other peers, etc... until eventually you get to the destination (which is the reverse of your connection)... The destination has their own ISP which has their own peers, etc.. somewhere they meet. There are only certain peers that your ISP contracts with, and conversely, there are only certain peers that the destination ISP contracts with. Through a bunch of magic routing, they eventually meet in the middle. Ultimately, it is the responsibility of your ISP to escalate peering issues until they get to the proper place, since carriers/backbones/etc don't deal directly with end-users. |
|
" On some cases, they do. I've had other issues in the past that I managed to fix by mailing the support of this companies. In some of this cases they don't care who complains, the issue exists no matter who is complaining. I mailed both companies with the MTRs, if the planets are aligned some guy that actually wants to do his job reads the mails and do something about it :p Oddly, compared to previous problems with the international routing and this particular ISP both OP and I have, this problem seem to have a 3-4 hour window. Previously it was more like a 12 hour window. IMPROVE! lol Last edited by daemmon#1312 on Nov 4, 2014, 8:54:43 PM
|
|
|
|------------------------------------------------------------------------------------------|
| WinMTR statistics | | Host - % | Sent | Recv | Best | Avrg | Wrst | Last | |------------------------------------------------|------|------|------|------|------|------| | 192.168.0.1 - 0 | 1478 | 1478 | 0 | 0 | 3 | 0 | | 1-197-28-181.fibertel.com.ar - 1 | 1462 | 1458 | 5 | 10 | 87 | 8 | | No response from host - 100 | 296 | 0 | 0 | 0 | 0 | 0 | | No response from host - 100 | 296 | 0 | 0 | 0 | 0 | 0 | | No response from host - 100 | 296 | 0 | 0 | 0 | 0 | 0 | | 117-165-89-200.fibertel.com.ar - 0 | 1477 | 1477 | 15 | 20 | 46 | 17 | | 197-165-89-200.fibertel.com.ar - 0 | 1477 | 1477 | 15 | 20 | 40 | 18 | | 222-165-89-200.fibertel.com.ar - 0 | 1477 | 1477 | 14 | 20 | 48 | 17 | | ge-0-1-2.ar3.eze1.gblx.net - 0 | 1477 | 1477 | 14 | 19 | 77 | 21 | | po2.ar3.MIA2.gblx.net - 0 | 1477 | 1477 | 139 | 177 | 1120 | 142 | | ae5.edge2.miami2.level3.net - 15 | 939 | 804 | 0 | 151 | 222 | 177 | | ae-2-52.edge1.Miami2.Level3.net - 24 | 759 | 579 | 141 | 153 | 228 | 172 | | te2-5.bbr01.tm01.mia01.networklayer.com - 0 | 1477 | 1477 | 145 | 149 | 192 | 148 | | ae7.bbr01.tm01.mia01.networklayer.com - 0 | 1477 | 1477 | 145 | 149 | 175 | 148 | | ae1.bbr01.sr02.hou02.networklayer.com - 1 | 1470 | 1468 | 232 | 269 | 299 | 242 | | ae3.bbr01.eq01.dal03.networklayer.com - 1 | 1458 | 1453 | 235 | 268 | 305 | 245 | | ae5.dar02.dal06.networklayer.com - 1 | 1474 | 1473 | 226 | 274 | 317 | 238 | | po2.fcr01.sr02.dal06.networklayer.com - 2 | 1418 | 1403 | 175 | 182 | 293 | 185 | |192.155.227.218-static.reverse.softlayer.com - 0 | 1477 | 1477 | 213 | 264 | 306 | 227 | |________________________________________________|______|______|______|______|______|______| WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider Problem seems to be starting on a later hop today. | ae7.bbr01.tm01.mia01.networklayer.com - 0 | 1477 | 1477 | 145 | 149 | 175 | 148 | | ae1.bbr01.sr02.hou02.networklayer.com - 1 | 1470 | 1468 | 232 | 269 | 299 | 242 | |
|
" What you are seeing is not a problem. The "packet loss" showing up is not real packet loss as you sent a total of 1478 packets from your router, and 1477 were sent/received by the server. The discrepancy of one packet might just be that you shut off the trace mid-ping. Really, I see your issue here: | ge-0-1-2.ar3.eze1.gblx.net - 0 | 1477 | 1477 | 14 | 19 | 77 | 21 | | po2.ar3.MIA2.gblx.net - 0 | 1477 | 1477 | 139 | 177 | 1120 | 142 | Notice it goes from 77ms worst ping to 1120ms worst ping? That kind of jump is bad. In addition to that, it goes from an AVERAGE ping of 19ms to an AVERAGE ping of 177ms. Again, a large jump in latency like that is not good and can point to peering issues. |
|
" Oh, I wasn't talking about the packet lost, I was talking about the 175ms average to 300ms average. Also, those 1120ms spikes are common, that's what you get when you try to shoot something from Argentina and grab it at Miami. That's extremely common on 70%+ of the connections done to USA and we get used to that kind of spikes. That's something that will only get fixed when we get BR/SA gateways. That 175ms to 300ms hop could be a problem with the routing I'm getting when the packet is returning (something my brother was telling me the other day, as he works for another ISP down here and has solved some issues with PoE) btw, when I'm outside the peak hours I still get those 1120ms spikes. I'll do a MTR when everything is working later to double check and post. |
|
|
|------------------------------------------------------------------------------------------|
| WinMTR statistics | | Host - % | Sent | Recv | Best | Avrg | Wrst | Last | |------------------------------------------------|------|------|------|------|------|------| | 192.168.0.1 - 0 | 1773 | 1773 | 0 | 0 | 2 | 0 | | 1-197-28-181.fibertel.com.ar - 1 | 1741 | 1733 | 5 | 10 | 80 | 10 | | No response from host - 100 | 355 | 0 | 0 | 0 | 0 | 0 | | No response from host - 100 | 355 | 0 | 0 | 0 | 0 | 0 | | No response from host - 100 | 355 | 0 | 0 | 0 | 0 | 0 | | 117-165-89-200.fibertel.com.ar - 0 | 1773 | 1773 | 15 | 21 | 46 | 21 | | 197-165-89-200.fibertel.com.ar - 0 | 1773 | 1773 | 14 | 20 | 44 | 20 | | 222-165-89-200.fibertel.com.ar - 0 | 1773 | 1773 | 15 | 20 | 45 | 22 | | ge-0-1-2.ar3.eze1.gblx.net - 0 | 1773 | 1773 | 14 | 19 | 97 | 17 | | po2.ar3.MIA2.gblx.net - 0 | 1771 | 1771 | 139 | 172 | 1944 | 152 | | ae5.edge2.miami2.level3.net - 8 | 1365 | 1263 | 0 | 151 | 243 | 156 | | ae-2-52.edge1.Miami2.Level3.net - 28 | 852 | 621 | 142 | 148 | 221 | 159 | | te2-5.bbr01.tm01.mia01.networklayer.com - 0 | 1772 | 1772 | 144 | 149 | 248 | 149 | | ae7.bbr01.tm01.mia01.networklayer.com - 0 | 1772 | 1772 | 145 | 149 | 253 | 148 | | ae1.bbr01.sr02.hou02.networklayer.com - 0 | 1772 | 1772 | 176 | 181 | 212 | 189 | | ae3.bbr01.eq01.dal03.networklayer.com - 0 | 1773 | 1773 | 176 | 180 | 232 | 182 | | ae5.dar02.dal06.networklayer.com - 0 | 1773 | 1773 | 188 | 192 | 223 | 191 | | po2.fcr01.sr02.dal06.networklayer.com - 1 | 1761 | 1758 | 175 | 186 | 521 | 179 | |192.155.227.218-static.reverse.softlayer.com - 0 | 1772 | 1772 | 175 | 182 | 211 | 188 | |________________________________________________|______|______|______|______|______|______| WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider This is how it works on playable status (same server as the previous MTR). Spikes still there, but as I said... they are common :) Edit: Just in case, as I said before, we are not intending to play on anything below 180-200ms. We know that's the best we can get to USA, lowest I've seen is 140 to certain specific servers on Florida, but that's just best routing possible + closest to our way of reaching USA. The problem starts heavily when hops on Networklayer go 280-300ms. That's totally unplayable for us. Last edited by daemmon#1312 on Nov 5, 2014, 9:42:52 PM
|
|
