I have been looking around for a few weeks now on what I should do with a spare server or two and came across Storj today. Got all pumped until I saw the proposal for the reduced payouts. Since it didnt happen yet I thought I would still try my luck and spinning up a node and adding my storage to the network, but was hoping to ask some questions. I found BrightenSilences calculator which I assume only works if you are already an operator so I went searching online for other calculators just to check if becoming a SNO would not be a negative thing for me but I couldnt find a solid answer.
I currently have 70TB of raw storage on one of my spares and my bw is 6Gb down/up. Running this thing costs a pretty good penny and I understand that my storage wont fill up instantly so I was curious what a new node could see in its first few months.
Does having cache drives help at all in the network?
Since the proposal is something I would be concerned about if I joined, would the higher end servers take a big hit?
My plan for Storj would be to colocate after truly knowing what my hardware can do. So if I were to move servers around or have downtime what actually happens? I see that you get a suspended if its a long period of time? Only downtime I see happening is my power going out at home. I can probably find this answer in the docs as I continue to research but thought I would ask to see if anyone has had any experience with it.
No, it actually takes the initial vetting period into account too. That’s the best tool you have. Remember though that it is based on historical data. Storj earnings depend a lot on how customers use the storage, and this is not possible to predict exactly.
With suboptimal setups they do. If your setup is tuned for Storj, they won’t. A lot depends then if you’re willing to spend time to learn how to make storage nodes perform well.
So far the proposal was presented as something that would affect all node operators, whether big or small. So yes, you should take it into account.
A lot of people here moved their nodes across hardware. There are procedures that aim to keep the downtime as short as possible. Few days of downtime was acceptable so far and will not affect your nodes much.
I made the mistake of making an earnings calculator and an earnings estimator with way too similar names. Keep in mind though that the estimator uses the current payouts which will change over the course of the year. I’ll try to keep things up to date as new information on payouts becomes available.
The current withheld system also means you won’t keep 100% of earnings until month 10. For the first three months at least your node will earn almost nothing since withheld is at 75% of earnings. So it is very unlikely you will cover your costs during that time at a minimum.
Thanks! While I was playing around with the sheet, I found some other docs for requirements. Is it true each node max’s out at 24TB? So if I have 70 raw TB, I would have to spread this out across a few different servers. With out the payouts changing, how does anyone break even for the cost of the electricity, internet and any maintenance on the nodes?
Spreading out 36 of my drives across 9 different servers I can get approx 41.2TiB of usable space after a Raid Z1 and thats only if I do not have any redundant servers. With the max potential, I only get about $135 according to the calc at current payouts, with electricity that might cover all 9 for my area.
The purpose of the withheld system is more to build up some credit to help pay for repair costs if you leave the network. (To my understanding it doesn’t pay the costs fully.) Since there is a low barrier for entry for a new node. There is the possibility this could be replaced or augmented by SNO’s paying a deposit fee but nothing is decided yet so it is just the wthheld system for now.
There is no hard limit how much space do you want to allocate, the reason is different: at some point the equilibrium between uploads and deletions become close to some number, at the moment it’s
So @penfold warned you about the held system for new nodes, not postponed payout, however they both can prevent you from receiving any money for a long time (unless you opt-in for zkSync and get 10% bonus, but it will help only with the Minimum Payout Threshold).
The theoretical limit with current values is about 88TB. You can easily see it by just filling in a high value like 1000TB. It’ll display a warning and show the 88TB under max potential as well. However, this month has had ingress double. If things remain like that, node filling speed is doubled and the max potential is as well. It’s less useful these days to think of the theoretical limit, but more useful to just look at the 10 predicted years to get an idea of how long it will take to fill up what you want to fill up. 41.2TB would take about 6 and a half years. Though as mentioned, this could be cut in half if the ingress amount we’re seeing this month becomes the new standard. I tend to be slow to incorporate temporary spikes into these numbers as they tend to be temporary. But I’ll raise them if it remains like this.
I saw the great ingress we’re having and no. of nodes pretty constant. So it seems the rise in ingress is coming from customers, not nodes leaving. Maybe is just a spike, but it’s realy great anyways. It’s making you think twice before deciding of leaving.
Storj already on 20PB, no one rising so fast like storj today. hope soon will be 100PB
Green looks suspicions, 10PB of free space at one moment, someone added datacenter?
The graph looks promising from a business point of view. It sort of seems like you really don’t want to run high powered machines for Storj but a raspberry pi NAS instead so that you barely use any electricity.
As I mentioned. I planned on starting a node and then depending on if I can cover my losses, deploy to a colocation and expand on storage. With the /24 limitation which is understandable if it only affects yourself and the pretty low ROI at current payouts, it really just seems like I would be paying Storj to give them my storage instead of the other way around.
If I can ask, how long have you all been running your node(s)? How much storage do you have in each? Have you seen any profit past expenses?
I am extremely interested, I am just trying to wrap my head around this especially since the payouts are decreasing. Someone with a mix of 10 Dell R720s and R820 and a few other models, I can provide a PB or two pretty quickly, with caching even though it seems like it’s not needed for storj right now I believe. I understand I can’t provide that since it’s just a waste of money since the theoretical limits are much lower.
Doing the ROI calculator on filecoin, even though it’s probably inaccurate so I’m taking that into account. If I provided 2PB of storage plus the other necessary hardware and taking into account the colocation price, the return is much higher than I would assume but there’s still a return calculated. My research on Storj is much higher than filecoin so I am probably misunderstanding something in the calculator that filecoin provides.
Again, just trying to understand. Appreciate everyone responding.
I have around 400TB of space and 265 of it is full. I have several locations for all that. I have been here around 42 Months. I build all this step by step. Today I get around 600-800$ a month.
So for me it is profitable. If you have 2 PB you can make nodes, and put chia to free space, and then slowly delete it when nodes are filing up.
Just to add to Vadims Info. They are one of the biggest Node holders that actively participate in the forum. Due to them also setting up multiple IPs in different /24 subnets the filling rate is a lot higher than if you had it all in one /24.
Mostly StorJ is set up as a decentralized Cloud that has lots of nodes all across the globe. For that reason the focus isn’t really on big Datacenters / Nodes and rather smaller ones scattered all across that run on Hardware that is already running and space that is already bought and just currently not needed.
So your big professional Server would have to try and compete with the lost cost Rasberry Pi Nodes running one 16TB disk and nothing else. Both you and the pi would (more or less) get the same amount of new data per month.
I would recommend just setting up a small Node somewhere as a docker with 1TB or so of space and just see how it all works. If you don’t like it its always easy to just delete the Node again.
And if you do plan to stick around longer you would have a node that is already on the network for a while and might have completed the vetting process and might have processed further along the Held Back Times where you would only get part of your earnings payed out.
That is if you can host the docker somewhere on a server thats already running anyways. If you later want to migrate the node to a different server thats fairly easy by just doing a 1to1 copy of all files to the new server.
I set up one node end of December 2022.
I have had quite a few problems at first due to DDOS protection decreasing my performance of the node without me directly realizing it right away.
For the first 23 Days of March I have made 7,57$ (6,03$ Download, 0,03$ Repair, 1,51$ Storage.
I am still in my 50% held back period so currently 3,79$ would actually be payed out.
The estimated actual payout is currently at 5,25$
For Comparison the whole of febuary was 4,65$ (4,11$ Download, 0,54$ Storage), Actual payout after HeldBack: 1,16$
currently have 1.95TB Stored Space, 60-65GB of new Data Ingress, 15-20GB Egress.
Do be Aware that lots of uploaded Data you get from Ingress can be deleted shortly after so your Node does not grow by those 60GB a day. And the more data you have the more will on average be deleted.
From the 4th to the 22nd of March my node grew 890 GB or ~50GB a day.
Thanks. I’ll probably stand up a 12TB RPi in that case. I love the idea of this but I really can’t see how people who have racks in a data center are paying for the electricity and circuit for the internet, unless they have been around since the beginning I guess.
Are you deployed in data centers? At least where I am, $600-$800 a month breaks even for one rack, and that doesn’t even include an internet circuit and any remote hands if needed.
