ZeroTier File Transfer Speed
-
@scottalanmiller said in ZeroTier File Transfer Speed:
@Dashrender said in ZeroTier File Transfer Speed:
@scottalanmiller said in ZeroTier File Transfer Speed:
@Dashrender said in ZeroTier File Transfer Speed:
Granted, in most cases you should get something pretty damned close to one of the generic testing sites as well - which we assume the OP did when he said he tested multiple time
No, you should not. That's a crazy conclusion. It's exactly the opposite... you can almost guaranteed not expect anything close to that speed.
huh - I frequently get within 5% of the speed - and more often than not, I get more speed than I paid for.
Across the internet? I find that implausible. There are SO many bottlenecks, no one gets anywhere close. You can't use something like Speedtest to test as that specifically avoids all of the real world bottlenecks and tests the speed of your ISP, not the speed of your point to point capacity.
OK maybe not - but that is what I was basing my comments off of.
-
Can you count a Google Speed test? or does that too avoid all the pitfalls?
-
@Dashrender said in ZeroTier File Transfer Speed:
Can you count a Google Speed test? or does that too avoid all the pitfalls?
No speed test can be used. You need to set up a point to point connection and test yourself.
-
What about setting up iperf on both endpoint?
https://iperf.fr/ -
@scottalanmiller said in ZeroTier File Transfer Speed:
@Dashrender said in ZeroTier File Transfer Speed:
Can you count a Google Speed test? or does that too avoid all the pitfalls?
No speed test can be used. You need to set up a point to point connection and test yourself.
Well of course, if the goal is to test between the two endpoints - which is what the OP here wants. But wasn't what I was claiming that I've tested (and frankly - when the OP claimed they tested their ISPs, I also assumed they simply tried a speedtest.net test as well - showing that it WAS possible on the open internet to get speeds as sold by the ISP - nothing more).
Now, of course, as you and I and others have all said - the OP needs to setup an unencrypted transfer between these two specific sites to test that specific data path on the internet to see what kind of throughput he can get.
-
@Dashrender said in ZeroTier File Transfer Speed:
Well of course, if the goal is to test between the two endpoints - which is what the OP here wants. But wasn't what I was claiming that I've tested (and frankly - when the OP claimed they tested their ISPs, I also assumed they simply tried a speedtest.net test as well - showing that it WAS possible on the open internet to get speeds as sold by the ISP - nothing more).
Those tests are not the open Internet, they are specifically avoiding that. That's the point of them. They detect which servers are not going to the Internet techically at all and are staying within a single carrier.
-
@scottalanmiller said in ZeroTier File Transfer Speed:
@Dashrender said in ZeroTier File Transfer Speed:
Well of course, if the goal is to test between the two endpoints - which is what the OP here wants. But wasn't what I was claiming that I've tested (and frankly - when the OP claimed they tested their ISPs, I also assumed they simply tried a speedtest.net test as well - showing that it WAS possible on the open internet to get speeds as sold by the ISP - nothing more).
Those tests are not the open Internet, they are specifically avoiding that. That's the point of them. They detect which servers are not going to the Internet techically at all and are staying within a single carrier.
In the DC test above, I specifically looked for an endpoint NOT on Cox.
Including Google?
-
@Dashrender said in ZeroTier File Transfer Speed:
@scottalanmiller said in ZeroTier File Transfer Speed:
@Dashrender said in ZeroTier File Transfer Speed:
Well of course, if the goal is to test between the two endpoints - which is what the OP here wants. But wasn't what I was claiming that I've tested (and frankly - when the OP claimed they tested their ISPs, I also assumed they simply tried a speedtest.net test as well - showing that it WAS possible on the open internet to get speeds as sold by the ISP - nothing more).
Those tests are not the open Internet, they are specifically avoiding that. That's the point of them. They detect which servers are not going to the Internet techically at all and are staying within a single carrier.
In the DC test above, I specifically looked for an endpoint NOT on Cox.
Including Google?
Google is specifically designed to do the same thing as SpeedTest, it's about testing YOUR speed, not the speed of your point to point. All things Google run on local networks, that's part of their magic. Their DNS, their searches, their everything is hosted on all major (and man minor) carriers to make sure that accessing Google stuff never traverses carrier boundaries.
-
This is how the internet works:
Internet is made up of a bunch of independent but interconnected networks.
The Tier 1 networks are the internet backbone. They have peering agreements between each other to exchange traffic without paying.
Tier 2 pays money to send traffic over Tier 1 and Tier 3 pays money to send traffic over Tier 2. And the end users pays money to send traffic over Tier 3.
The bandwidth sold to the users are seriously oversubscribed.
So when traffic is moving over the internet it passes a bunch of different networks from source to destination. All of these have bandwidth limitations just like any physical network and there is a lot of traffic. As you pass more networks the latency increases and the bandwidth decreases. How much depends on whatever else is going on at the same time.
Big companies like Google have their own network infrastructure with their own fiber backbone so they can bypass some of the internet network infrastructure.
If you have two servers connected over the internet you are wise to use the same ISP because the traffic is more likely to stay on the same network. And the closer they are physically, the better it is.
If you however are physically close but on different networks, the traffic might have to traverse many networks before getting to it's destination.
So the actual bandwidth and latency you end up getting between server A and server B is a crapshoot.
-
@black3dynamite said in ZeroTier File Transfer Speed:
What about setting up iperf on both endpoint?
https://iperf.fr/This is one of the best solutions.
-
@JaredBusch said in ZeroTier File Transfer Speed:
@black3dynamite said in ZeroTier File Transfer Speed:
What about setting up iperf on both endpoint?
https://iperf.fr/This is one of the best solutions.
It's just a method for finding the answer to the question that's been asked many times already - what is the speed when transferring point to point between these two locations.
-
@Dashrender said in ZeroTier File Transfer Speed:
@JaredBusch said in ZeroTier File Transfer Speed:
@black3dynamite said in ZeroTier File Transfer Speed:
What about setting up iperf on both endpoint?
https://iperf.fr/This is one of the best solutions.
It's just a method for finding the answer to the question that's been asked many times already - what is the speed when transferring point to point between these two locations.
WTF? It is the answer to the OP's question. fuck the sideways damned discussion.
The OP and open up some port forwards and run iperf between two computers.That will prove his line speed without VPN.
-
Conveniently, he can also do the same over the ZT network so that he has speed tests from the same tool on both networks.
