I found some good preforming nodes on the same subnet (not mine). I noticed that there are 20 online nodes on the same EU IP, and each node receives more than 80GB of ingress data per day. All of them are also pointed to the same wallet.
Now I’m wondering if this is normal behavior for older nodes on same subnet?
Most of the nodes I monitor usually get around 10-20GB of ingress per day - and in those cases, there is only 1 node per subnet. The more nodes there are on the same subnet, the less traffic each node usually gets.
Based on my results, I see a daily average of about 1 Mbps ingress per single node. But this user somehow handles a steady 7–8 Mbps per node (around 1.6 TB per day across 20 nodes).
My main questions are:
Did the user with 20 nodes on the same subnet apply some kind of NVMe optimizations for faster response times?
Do nodes that are 1+ years old receive more data compared to vetted nodes that have only been online for about 4 months?
Does having an online score of 100% result in more ingress?
Would nodes perform better on a 5/5 Gbps network compared to a 0.1/1 Gbps network, and in that case receive more data?
Does US node location preform better than EU location?
Well, having a faster node in terms of latency would indeed help. And, uh, 7-8 Mbps is plausible (this is around 1MB/s). But not 20×7-8 Mbps within a single /24 block.
I am also curious as to how you measure their traffic (-:
All other factors equal, no.
Your node can only receive ingress if it is online, so I would say, yes (-:
That said, if you often have brief offline periods, this will keep impacting success rate, and therefore power-of-two-choices algorithm will be choosing your nodes less often.
All other factors equal, only if you’re saturating your network.
Are you maybe looking at a Select install? They don’t have /24 restrictions, and we know there are customers like Vivint pouring a firehose of data into it.
I found those nodes through Google because their dashboard ports are exposed on public IPs and indexed in Google search.
At first, I was skeptical that they were all on the same IP, so I ran a soft port scan to verify that the nodes are indeed running there (to rule out the possibility that the IP is used only for statistics).
There are 20 dashboard ports and 20 main node communication ports.
Location is Finland, it’s a datacenter IP address (maybe running in datacenter or IP is just acting like a VPN tunnel).
Forgot to mention that this user operates more than 20 nodes for sure (assuming 20 per subnet), because the monthly payment to the SNO wallet address is stable around $2,500.
Could still be that traffic to the nodes are proxied via other addresses/VPS/…
You can’t see what host:port the node registers with from the dashboard.
However, such a “simple/messy” setup would not surprise me if the same operator forgot to close the console ports.
)
