Bandwidth limitations and Node-per-ip

Strange. I see that both you and @tankmann are both close to around 1TB total bandwidth usage for the month. I’m seeing around 750GB~ for three nodes behind one ip address. Are others around 1TB so far as well? Wondering if there is something off with the “node-per-ip” limitations?

818 GB so far…

1.13 TB here.

About 1TB every month

hi @ndragun
Jocelyn here - community manager :slight_smile:
I split your comments off from the original conversation as I thought the topic might merit its own discussion and to keep the original thread on-topic. Hope you wont mind – also, I was typing in haste, so if you have an idea for a better subject line, just ping me and I’ll change it

Please everyone carry on with your discussion!

There is limitation, avery node behind 1 IP is looks like 1 node, if you have 5 nodes behind 1 IP, network will give you data as to 1 node together for all of them.

3 nodes behind the same ip, 600GB out, 500GB in

Is that a total of? Is it for each node?

it’s the total of all 3 nodes

Interesting :confused:

266GB / 265GB for the whole month for one Node.
(Industrial Hardware on Gbit uptime >99.9%)…

Looks like different node priority based on geographical location
and Nodes in ISP Subnet??
Thought there is no classification?

1 Like

2.2Tb of egress on 5 nodes/5ip/3 location

where is a couple of new nodes, they do not yet make a significant contribution to statistics

There’s not really a “classification” so to speak… However, if two nodes identical nodes (node speed, as well as Internet connection speed) are situated at geographically distant points from a given satellite, the farther node will receive fewer data pieces… because signals can travel no faster than 3x10^8 m/s … So, the closer node will win the speed race most of the time even though both nodes are “identical” in build specs. There’s no way to fix this if centralized satellites are the distribution points for data pieces.

This was my expectation as well, but I just set up a new node this week because the other was approaching 75% full, and the new node has about half of the traffic as the old node, but the old node’s traffic pattern has not measurably changed.

I run three nodes now, one is at a different location. That node shows the same traffic pattern in general, so I know that there wasn’t a coincidental increase in traffic across the network at the same moment I launched my new node. Rather, the network does appear to be sending more traffic to this IP since launching a second node there – a 50% increase in traffic by my estimation.

It’s odd for sure… I’m not complaining as an SNO, but for the health of the network I am curious what’s going on and if there’s an issue.

The new node is not veeted yet? Thin it will defenetly geting less data, untill veeting will complete.
it should be 100 audition on every satellite.

Yes, the new node is not vetted. However, it’s getting a pretty significant amount of ingress traffic – more than I would expect for an unvetted node, and the other node at the same IP doesn’t appear to be getting less traffic.

The satellite doesn’t distribute data, it distribute nodes list to uplinks, the uplink connecting directly to the nodes and upload/download data.
So, distance to satellite doesn’t matter, but distance to customer it is.

1 Like

Could you explain how it works?

If customers uploading a file, the satellite will send a list of uplinks from nodes so the
customer will send an upload request to “x” nodes and the “race” starts?

The fastest “x” nodes will download the file, all slower ones will fail this request.
Am i right that the satellite will send the list independent of the distance from customer to node?

Or does the node separate the nearest nodes from the list?

If customer want to upload a file, their uplink will request a list of nodes from the configured satellite.
Then uplink do all magic with encrypting and splitting the file, then starts uploads to nodes in parallel. When the number of successful uploads is matched the configured threshold, it cancel remaining uploads.

When customer want to download a file their uplink requesting satellite for list of nodes where pieces of the file are located and starts downloads in parallel. When the number of successful downloads is matched the configured threshold, it cancel remaining downloads and assembly the file and decrypt it.

You can read details here:

Okay got it.

Reputation is the Key…

Currently yes, it’s random. The uplink just request “Hey, give me 130 nodes!”
I think it would be improved over the time.

1 Like

A post was split to a new topic: Is there a way to check and compare your reputation of nodes?