here are my stats from yesterday (01Nov2020):
3 Nodes:
Nodes 1 and 2 are vetted and each received about 18.2 GB each, and then Node 3 is going through vetting now and got 4.3 GB ingress.
Hmm my numbers do seem slightly off… but i do have 1 or 2 nodes that are vetting… maybe that affects it somehow…
thanks for the quick responses, i guess ill have to go through my stuff… yes i am talking about ingress only… my egress is pretty good
for the 1st of nov my egress was slightly above 58GB on my 14.5TB node, been undulating between like 40gb and 80gb egress a day for a long time now, think my oct total was 2tb egress + repair egress
I had ~37GB ingress and ~10GB egress (for 4TB stored)
i did have some stuff running that was using 10mbit up and down or more globally and running 9000 connections tho only 1700 active
maybe that affected my storagenodes, but my successrates have been 99.9 or 99.8% on up and down for the last few weeks.
maybe that latency thing they talked about adding is actually working… sure seems like it maybe… i tried turning my other services off.
25GB vs 36GB is still pretty far off… you are literally getting 50% more
anyone that know how vetting affects ingress? when having multiple nodes on the same ip?
i mean if a vetting node takes up a share of the ingress without actually getting the ingress, then my numbers fit… i lack about 1/3 and i got 1 old node and 2 new nodes running where i think 1 is vetted.
so if a vetting node takes its hare to the ingress, but without getting the actual ingress, then my numbers makes sense.
1 Like
For November 2nd:
Ingress: between 35 - 40GB per node. (with repair)
Egress: between 18 - 27GB per node. (also with repair)
You could add up all traffic from all nodes that you have and it should end up ~37GB in total (given that you are comparing our data from first of november)- that’s what they mean when they say about traffic distribution per IP.
When my secondary node was during vetting process - my summed up traffic was pretty close to what you were getting from one node.
Hope it helps.
it seems like it’s 1/3 of the max ingress i lack and funny enough i got a node vetting… so seems the simplest answer, but i ofc duno really… but i will test it in the future… it would represent a substantial loss if it’s the case. ofc even with 3 nodes and then say 1 month to vet, on a yearly basis it would only be a few 3% lower
and as one starts to have higher number of nodes, and lets say just vetting one at a time… then it would be a fraction… so like say 10 nodes its 10% losses of ingress for a month…
so not a huge deal… but something to be aware of i guess… maybe 
and it’s not just close… i got like 25.2 GB ingress… and if that is 2/3rd’s then it would be 12.5gb added so 37.8, so basically spot on… so logic seems to dictate it has to be the reason…
no, there is no connection… but you can use the search. vetting distribution etc has been linked multiple times in the last couple weeks.
well you are right… your argument made me go back and scan my subnet because if i had 2 nodes active and 1/3 was missing… that would mean there should be an extra node leeching ingress…
i swear he was gone for a long time… days … atleast… i suppose i should have checked a bit more frequently…
That is very strange unless you are not in charge of your subnet router - I mean you have to forward ports on your router to get traffic to a node, right?
Are you the only person with access to the router in your subnet? Does someone else have access?
Is there any other possibility to create a traffic to node, without port forwarding? (except multiple public IPs)
the subnet is controlled by the ISP leasing out IP addresses in the, ip/24 subnet of which i’m leasing 1 ip address, just like most other regular consumers if the lease a global ip.
then ofc because they apperently are filling 1 subnet at a time and that there is somebody which got the same idea of setting up a Storj node and leased an ip and then was shuffled into the same subnet as me…
small high tech country problems, something that i would expect the rest of the network will have to deal with also when it becomes more popular… but i’m switching ISP… i guess i should get around to get that project finished… so i can get back to 100% ingress again.
2 Likes
So… do you have public IP (dynamic or not - doesn’t matter) ? or did you have to ask to forward a port to your leased subnet IP? (does your router IP start with 10.* or 172.* or 192.* )?
Because if you have public IP then no one outside of your house router can share your STORj traffic…
But as you say - you are due to ISP change, maybe it will be for the better…
no… traffic is shared among the PUBLIC /24 subnet of an ip.
1 Like
I think that public IP means that your router has one of a kind address that is unique in the whole world.
So one public IP gets one “unit” of traffic - no matter if you have 1, 20 or 100 nodes in this subnet of this one public IP.
If you have public IP, you should get “full” traffic to this IP. Whether you decide to split it into 1 node, 5 nodes, that means that 1 “unit” of traffic will be split to 1 or 5 nodes.
By “unit” I mean “daily traffic” so it may vary depending on a day - once it can be 37GB next day maybe different value. I used it as simplification of daily transfer.
Only thing that matters is if you have public IP or not.
If your router’s IP (on WAN port of your router) starts with 10.* or 172.* or 192.* then your router doesn’t have public IP - that means you may be sharing STORj traffic.
And you probably had to ask ISP to forward traffic on one of the ports from ISP’s public IP so you will get STORj traffic, or you have use some kind of trick to get STORj traffic.
You and SGC like to think/speculate a lot, but please please check your facts and use the search instead of spreading your thoughts/speculations! That only leads to confused people and to other forum members having to correct people who seem to read every thread and can be considered regular users…
Just one of these threads you can find using the search is this:
You share traffic with everyone in the PUBLIC /24 subnet of your routers public IP. That’s how storj works to geographically distribute the data. So you can be in the same subnet as your neighbour and therefore share traffic with him.
yeah because if you have no public IP, you can’t even forward the storj port and therefore not participate in the storj network at all… You can only do so with a public IP, whether it’s your router that got one or you use a proxy/vpn.
5 Likes
the problem is that the satellites doesn’t use the ip address to select where the data goes, they select a subnet aka IP/24 and thus groups of 254 ip addresses are selected as one and any nodes within that group will have the data split between them.
there isn’t any room for making an argument about it, it’s mechanics and almost open source… not sure if they actually have the latest code for it tho… .because they don’t want people gaming the system to much.
but still it’s not secret how it works atleast the basics of it and from my understanding it is a sort of a stop gap measure because they don’t have a better solution for the problem posed by how to distribute the data… alas it doesn’t matter because it’s just what we have to live with for the foreseeable future.
and yeah it doesn’t really have anything to do with lan ip addresses or port forwarding…
that’s basically totally irrelevant to the discussion
2 Likes
I knew that (but ignored that 
as I thought it is pretty lazy and unfair for different users) - it didn’t make any sense for me until now, when I “got” it… sorry for spreading wrong info to people.
That is pretty weak assumption made by STORj.
I have an ISP that is the biggest in my country and he supplies dynamic public IPs. One day I have an IP that can be allocated to totally different part of the country the very next day. I’ve noticed that one day various “what’s my IP” services locate me in different parts of the country (even more than 200km away from my location).
This ISP is country-wide and it’s service covers an approximation of rectangle that is ~800km tall and wide.
So pretty unfortunate assumption for “data” distribution. At least for this ISP 
That also means that each day I may get “lucky” and share a /24 subnet with someone who decided to enter STORj project, and who lives half country away from me
That also mean that someone can have a STORj service in the same building as me and have IP that is geolocated hundreds of kms from this building…
yeah nobody will argue against it being not a perfect solution but apparently it’s the best storjlabs could come up with and I haven’t heard any better proposals either yet.
My geolocation can be very far away too but in over a year I have never shared a subnet with anyone, even though most nodes are located in Germany. But I have a fairly “stable” ip, maybe it’s different for people using dsl but haven’t heard many complain about sharing traffic. It’s just a few unfortunate souls (and datacenters obviously, at least the cheap ones)
2 Likes