New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
Really comes down to the upstream and what route at the given time it wants to take.
There is really never a one best location to everywhere.
Chicago and Dallas for a long time was a major spot for game servers to get a central latency. Ashburn as MikeA said is a solid option as well.
I remember years back with our Seattle location traffic to Vancouver BC was rock solid super low latency, then routes started going to California and then back until after a month or two GTT fixed their routes. We see this happen all the time with different locations and routes due to fiber cuts or bad peering etc.
Oregon to Chicago on Comcast = 63-66ms
Oregon to New York on Comcast = 76-77ms
We found our Seattle location to Oregon AWS was 6.3ms over direct peering with AWS. When testing multiple providers in Oregon to AWS which should be 1-3ms was 10-13ms as all of the traffic for some reason would route up to Seattle then back.
@PureVoltage
Thanks a lot for that really useful (and real-world) information! Something like that is well in the range of what I hoped for
Its where nsa has its catch all internet traffic datacenters
Where is "it"?
ashburn of course
What does this mean? I just started as relatively new intern at one of the largest cloud platforms, nobody has Seattle location (which is honestly a little strange given how AWS is based there too).
AWS, G both northwest primary is Oregon (AWS have a local but not reall zone for customers), MS have one that is 4 hours drive away.. GPT just inventing nonsense
Azure is the only one of the big three cloud providers with presence there. Hetzner and OVH both added Oregon a bit ago.
There is a lot of connectivity from and into Seattle.
Fiber cables coming in from Asian markets, Japan, Hong Kong, Korea. Some of which flow into Oregon but goes directly up to Seattle.
Also links up a lot of Canada on the West coast and Alaska. Quite a lot of West <-> East traffic flows
For the cloud providers it's a funky one because a large amount of anything in Oregon still runs into Seattle. This was some tests we did a little while back for a client wanting quick connections to AWS in Oregon turned out Seattle got better latency for them at 6.3ms
MTR's to AWS for example.
Success rate is 100 percent (5/5), round-trip min/avg/max = 14/14/14 ms
Thu Jun 12 17:39:59.804 UTC
Type escape sequence to abort.
Tracing the route to 44.248.68.30
1 be17.cr3-sea-b.bb.as11404.net (174.127.136.154) [MPLS: Label 24590 Exp 0] 5 msec 5 msec
2 be12.cr6-sea.bb.as11404.net (174.127.151.12) 5 msec 5 msec
3 * *
4 108.166.228.9 9 msec
5 * *
6 108.166.232.33 10 msec *
7 * *
8 ec2-44-248-68-30.us-west-2.compute.amazonaws.com (44.248.68.30) 11 msec 12 msec
IPv4 ping to 44.248.68.30
PING 44.248.68.30 (44.248.68.30): 56 data bytes
64 bytes from 44.248.68.30: icmp_seq=0 ttl=57 time=8.864 ms
64 bytes from 44.248.68.30: icmp_seq=1 ttl=57 time=9.030 ms
64 bytes from 44.248.68.30: icmp_seq=2 ttl=57 time=8.657 ms
64 bytes from 44.248.68.30: icmp_seq=3 ttl=57 time=8.694 ms
64 bytes from 44.248.68.30: icmp_seq=4 ttl=57 time=11.750 ms
--- 44.248.68.30 ping statistics ---
5 packets transmitted, 5 packets received, 0% packet loss
round-trip min/avg/max/stddev = 8.657/9.399/11.750/1.183 ms
~~~~~~
^^^^ GTT Seattle ^^^^
~~~~~
PORTLAND OR USA Ping results for: 44.248.68.30 (ec2-44-248-68-30.us-west-2.compute.amazonaws.com)
Target: 44.248.68.30 with 64 data bytes Success rate sent/received = (5/5) packets
Round-Trip min/avg/max = 9ms /11ms /15ms
PORTLAND OR USA Traceroute results for: 44.248.68.30 (ec2-44-248-68-30.us-west-2.compute.amazonaws.com)
Tracing route to 44.248.68.30 1 ae4.6.bar3.Portland1.net.lumen.tech (4.69.220.210) 3ms 1ms 8ms
2 * * *
3 108.166.240.0 (108.166.240.0) 10ms 108.166.240.1 (108.166.240.1) 11ms 108.166.240.0 (108.166.240.0) 10ms
4 * * *
5 108.166.240.14 (108.166.240.14) 9ms * 108.166.232.70 (108.166.232.70) 11ms
6 * * *
7 ec2-44-248-68-30.us-west-2.compute.amazonaws.com (44.248.68.30) 12ms 9ms 11ms
Router: Hillsboro, OR
Command: ping count 5 44.248.68.30
PING 44.248.68.30 (44.248.68.30): 56 data bytes
64 bytes from 44.248.68.30: icmp_seq=0 ttl=56 time=11.328 ms
64 bytes from 44.248.68.30: icmp_seq=1 ttl=56 time=12.023 ms
64 bytes from 44.248.68.30: icmp_seq=2 ttl=56 time=11.183 ms
64 bytes from 44.248.68.30: icmp_seq=3 ttl=56 time=11.210 ms
64 bytes from 44.248.68.30: icmp_seq=4 ttl=56 time=11.190 ms
--- 44.248.68.30 ping statistics ---
5 packets transmitted, 5 packets received, 0% packet loss
round-trip min/avg/max/stddev = 11.183/11.387/12.023/0.322 ms
traceroute to 44.248.68.30 (44.248.68.30), 30 hops max, 52 byte packets
1 ae3.mpr4.pdx1.us.zip.zayo.com (64.125.20.222) 2.270 ms 2.024 ms 2.068 ms
2 ae6.cr1.sea1.us.zip.zayo.com (64.125.26.198) 4.938 ms 5.131 ms 4.896 ms
3 ae2.mcr1.sea1.us.zip.zayo.com (64.125.24.171) 4.883 ms 5.242 ms 5.121 ms
4 * * *
5 * * *
6 * * *
7 ec2-44-248-68-30.us-west-2.compute.amazonaws.com (44.248.68.30) 13.200 ms * 11.045 ms
What I meant was more how did GPT decide that multiple major clouds had locations there when none do or even claim to anywhere in documentation haha
That is interesting though, I've seen a bunch of clients end up running in Portland because it was lower latency than going to Microsoft's WestUs2 in Moses Lake/Quincy in the same state
Missed the GPT part there
A lot of those places are funky with the routes. We almost did a massive pop in Quincy just to offer ultra budget systems vs selling off / tossing / giving away old systems. However the connections there kept us from pulling the trigger.
I had a post years back on WHT about OVH's network basically all just flowing traffic from Oregon to Seattle back to Oregon. After doing a test a little while back seems it's still sadly happening.
GPT sometimes doesn't quite nail it on facts it has.
Read this part again? "I dont know about Internationally, but as someone living in Asia, my best ping to US is with New York"
Do you have any maps of the fiber runs that run from Asia to East USA? Because that makes no sense.
That would be news to me. Any real world examples? Toronto yes, Vancouver no. Telus and Shaw have sued each other on their fiber claims across Canada.
Yeah, OVH Oregon either routes via Seattle or Chicago.
You misunderstood. The part I quoted is the statement you made with respect to the internet. @raindog308 tried to make sense of what you were saying, because it was also unclear to him, and I pointed out that was contrary to the criteria you set in the OP so NOT what you meant. (On re-read, I didn't call out the backhaul as being contrary to "best connectivity" and bandwidth like I thought was obvious).
Of course, rather than clarify what you actually meant, you are butthurt and throw a hissy fit.
I think raindog will agree he's not a network guy and wouldn't claim to be a professional in networking.
But regardless, the issue is your vague and incorrect statement that I quoted in the first place for which you didn't clarify and instead whined like a bitch.
Dude, let's be honest: @jsg could post a five-line shell script that cured cancer and you'd castigate him for it.
If I didn't know you two are married in real life, I'd think you had it out for him or something.
That's disappointing. That tells me you don't know how full of shit he actually is.
But do give me some examples where I point out something that isn't actually wrong. I feel like I explain my argument every single time.
You must have like at least three examples if you're going to argue I'd be against the cure for cancer just because jsg made it.
Let's cut this short: you have oftentimes and consistently stated - publicly! - that you consider me clueless, full of shit, not even knowing what I'm talking about, and so on, and so on.
Now a simple question to you: What qualification do you have to judge technical matters AT ALL?
You never ever provided a shred of evidence suggesting that you are even remotely in a position to judge anything technical related to software, hardware, or network engineering. All you offered was blank assertions and judgements.
Actually it's you who is disappointing. And @raindog308 is right; I could make it rain gold, I could cure cancer or create happiness on this planet - with one exception: you. You'd still attack me, and be it because I spelled something wrong.
But still, I thank you. Because now after lecturing and judging even a deservedly well respected admin, maybe they get it and finally react adequately to your destructive, utterly negative, and condescending behaviour.
Alone for daring to condescendingly lecture raindog you should get a severe warning IMO.
Why? Because you say so? Hardly.
https://lowendtalk.com/discussion/comment/4289763/#Comment_4289763
That's a lie. Not even a few months ago (edit: Feb 7) I gave you lots on my background and qualifications and then I asked you for the same and you fucking whined and crawled away. You have never stated your background or qualifications. Fucking idiot.
The dozen times I've told you that you're spelling "loser" wrong, not once was insulting. But you take it many extra steps and intentionally spell things wrong, for unknown reasons.
I didn't lecture him. But I then judged him, and not for the first time.
Condescending lecture to raindog? GTFO whiny bitch.
To back up his assertion that I would attack you regardless of the content or merits. I explicitly stated I give reasons for why I call out your shit and wanted to know where that wasn't deserved. Duh.
You should consider seeing a therapist
Kindly point me to where you did that. And, well noted, I wasn't talking of assertions by you about yourself and/or your qualifications. I was talking about something tangible.
That's not for you, at least not for you alone to decide, but hey, I give you the benefit of doubt.
So, you know my intentions? You can look into my mind and my heart?
That may be, how you see it but you are not the final reference and everybody, incl. myself, is entitled to their own perception and view.
Besides vulgar language, so typical for you, and what many would perceive as an insult ..
Yes, that's how I perceived it.
Obviously it never crossed your mind that your feeling of needing to "call out [my] shit" is neither the only, nor the final criterion. Nor does your feeling "it's deserved" constitute a valid, objective, and justifiable reason or justification.
Obviously you also have a severe lack re. manners and social interaction.
But be that as it may, I have nothing to discuss with you, this responding to you was merely out of courtesy (pardon my use of a term alien to you).
Sooner or later one of us will (or will have to) leave. I'll leave it to the LET team to determine whose loss would be worse and/or more regrettable for our community.
Think of me as you please, even feel free to hate me, that's fine with me. All I want is an end to the damage you create here and an end of your amok run, preferably by your own insight. But I admit, I won't hold my breath.
@jsg and I could not be further apart on several things. For example, the war in Ukraine and anthropogenic global warming (I think, based on a comment I think I'm remembering right). But I also respect him on some technical areas and find his viewpoint interesting.
I'd say the same thing about you. As I recall, you're an atheist, which is 180 degrees from my viewpoint. But I find some of your comments interesting and your zingers often entertaining.
I just don't feel the need to constantly pick fights.
Anyway, that's my perception.
Most connections from India to US that I know of go through Europe and atlantic. Singapore is a bit more eastward so maybe a different route.
Based on your previous comments, you seem to have no clue how it works. You claim geography is somehow directly related to latency.
That said, Asia is big. South West Asia mostly routes through Europe to East coast US. Naturally, the latency goes to shit(230-258) if I connect to West Coast.
every city is good until some shark bite the cable
Interesting. Usually south of India is routed via Chennai->SG->JP->LAX/SEA/OR
You can get latencies as low as 195ms from Chennai to Los Angeles via Airtel.
You can get latencies as low as 195ms from Chennai to Los Angeles via Airtel.
I always found my connection going through TATA's EU pops.
TIL about Chennai's route. Which ISP?
Mine is a local ISP - downstream of Airtel and TATA.
But US West, almost 90% of the time is through Airtel and exits via Chennai.
I have noticed some East coast connections routed via EU, and thats always 240ms+ for me (mix of both Airtel and TATA transit)
Oh, good point, yeah I guess from India maybe east cost US via Europe could well be better. I've had a lot more focus on China in recent years, so I tend to only think about east Asia and forget about the huge amount of Asia to the west!
The sexual tension couldn't be this intense if they were married, I think you're mistaken.
Yep, that's also what I see in my benchmarks. Up to and including India typically routes via the east coast and further east that is, Singapore to Japan and usually also to Ozzyland is routed via the west-coast and the Pacific.
Yep, I just looked at my east coast WDC reference. To Mumbai, via Europe is about 200 ms. +70ms to Chennai seems to make sense, no matter whether via submarine link or through India itself (assuming they do have some decent fibers there).
But I'd welcome some more responses to my question ("within the USA").