Realistic earnings estimator

The way I see it is:

  • Tardigrade pretty much has enough nodes for now, so they don’t really need to work on collecting more. Hence they don’t care as much regarding the accuracy of the calculator.
  • The statement on preferring more small nodes is just a general preference, not a specific requirement. It’s better to have 100 nodes than 10 nodes, and it’s better to have 1000 than 100 nodes. But it’s just not as big improvement to jump from 1000 to 10000 nodes.

There are likely more important things to work on to get Tardigrade popular, and it would make sense from the business point of view.

The average amount of revenue per TB stored is going to be the same for two nodes that offer different amount of free space but are otherwise very similar. Data intake speed will also be the same for these two nodes.
Nothing wrong with bigger nodes being potentially able to utilize the space as long as there’s demand to store data.

Yes. But according to official earning estimator it should be ~ 7% difference in earning (2 864 ÷ 2 670) between 20 and 3 TB storage. And we clearly see that it doesn’t.

Official estimator represents their business model: a lot of distributed little (may be 3 - 5 TB ?) nodes for better reliability. It make sense, it sounds like a strategy. But in practice there is no limit. What bothers me is the violation of their own strategy.

Certainly it is possible to say that “STORJ has flexible strategy …”. But let’s be honest. Official estimator shows values that is 4 times bigger then in practice (at least for first month). It means that official estimator doesn’t represent any strategy.

Since Storjlabs makes money on all data stored and all egress, I doubt their business model depends on income estimates of individual nodes.

This is quite an assumption to begin with and I’m pretty certain that is simply not the case.

These sizes were clearly never the only target group. In fact the recommended size has always been between 8TB and 24TB. But all sizes from 500GB and up are supported. The network has no problem incorporating nodes of all sizes, but individual node operators may find it’s quite useless to assign more than 24TB as that would take a long time to fill up.

So I’m not convinced that the current traffic we’re seeing is a deviation from the initial strategy to begin with. And then there is the next point…

Storjlabs is a startup that just launched their first production ready product a few months ago. Being flexible with regards to strategy is not just possible, it is vital to survival in this stage. You roll with the punches and adapt to customer demand. It’s quite unique to actually be making money in this stage to begin with and I’m not saying Storjlabs is profitable as a whole (I don’t have those numbers), but their product is. By taking a margin on everything, they have a profit stream from every customer from day one. So that speaks to financial viability of their business model.

Now this was quite a deviation, just to say that I don’t think the inaccurate earnings calculator is a sign of their strategy not working. But it remains bad information to give prospective SNOs, because it implies the wrong advise for their setup.


I’ve made some changes to the sheet. I raised the ingress amount and put it in a separate parameter. It seems my nodes (same IP) get more than 2TB ingress this month total.

More importantly, I added deletes as a % from the total amount stored. This may be a little controversial, but it seemed necessary to show something important. At some point your node is likely to see an amount of deletes that equals the incoming traffic. This will be a sort of equilibrium and the amount stored won’t really increase anymore. I also changed the full potential line to show the numbers of either your full node or this equilibrium level, based on which is smaller.

The deletes are never as predictable as this suggests, so your node may get bigger than the calculated amount for a few months and then drop down below it. Hopefully it averages out to something close to this prediction. For now I set deletes to 5% of data stored. This is about what I’ve actually seen on my node on average since the last network wipe. This puts that equilibrium at 40TB. So for now, I’d say larger nodes than 40TB may not really be worth it. The sheet will now reflect this.


There are natural limits for a really large nodes.
You forget that there is not just constant income (ingress) of data to the node. But also some data deletes happen all the time. Storj is not mining, it is real data storage networks. And after data uploaded to node no longer needed by it’s owner(who pays for the storage) - satellite deletes it from nodes.
And more data node stores already - more it loose due to deletions every day/month as average number of deletes in direct proportion to the total amount of data stored.

This result in “filling speed” is not even close for being a constant. Data ingress may be about constant indeed (assuming customer demand uploading files and total node capacity over whole network stays at same balance point) although it fluctuate a LOT around avarage, but “filling speed” will decrease in some proportion of data volume node is already accumulated.
Because it consists not only of the positive incoming stream (which does not depend on the size of the node at all), but also of the negative stream (which grows with the size of the node).

There were also artificial limits on node sizes before (first 2 TB/per node, later 4 TB/node). They were either canceled or simply raised to higher values just few month ago. It’s possible that they are still active and after reaching a certain volume limit (say 24 TB currently indicated as the maximum recommended volume for example), the node will stop receiving new data.

Opps, like like @BrightSilence already explained point about “negative data stream” on nodes due to files deletes in more detail. I was answering before read topic to the very end.

A little bump up on this. We are getting more posts from people who were being mislead by the official estimator. Some stay, some leave, but generally it will only harm the reputation of STORJlabs if the official calculator stays this inaccurate.


I believe it should be removed and brought back when test data is no longer being used. Test data is so erratic and not a representation of real traffic which will also mislead SNOs thinking that it is sustainable, which is not.

1 Like

your assumption of testdata is misleading. It is an integral part of the network not only for testing but to ensure a consistent payout of SNOs. Not counting testdata would make calculations a lot lower and then it would be unrealistically low. You can’t have a realistic estimator if you exclude something.

And the test data is not more erratic than client behaviour.

We need a realistic estimator for the foreseeable future, which includes test data until there is enough client traffic making test data irrelevant.


Currently daily ingress/egress on test data is about 10-20x more than real data. Test data is not guaranteed or it should not be something that we should continue to expect in the foreseeable future. If you rework the calculator with test data and Storj decides to stop giving out test data, we are back to square one. SNOs making posts on the forms saying that the payout is too low.

This is why we should not even provide an official calculator and let SNOs do research themselves to find BrightSilence’s calculator to make their own judgement. That way SNOs are not given false hope on something that is not certain from an official source.

1 Like

Storj Labs had already said they try to replicate real customer behavior with test data. As long as they stick to that then test data just simulates more customers. That’s not the problem. By far the biggest problem is that the official estimator pretends the upload will use up to 50mbit upload constantly no matter how much space you share. I think SNOs should have something to have some indication of what realistic earnings will be. Even if large part of that is currently test data. I rebuild this estimator to be easy to update with new values if the network behavior changes. They could do something similar for the official one and keep it up to date with very little effort.

PS. I’ve received more requests for edit rights on this sheet. I’ve decided a while ago that every time I get such a request the size of the top message stating that I won’t grant those goes up by one point. Please read the message and copy it to your own account if you want to enter your own values.


I made some changes to this suggestion, since the original earnings calculator has been removed. Consider this suggestion a suggestion to bring a realistic one back. In the mean time this community estimator can perhaps help people make a good estimation. Since most people coming here are looking for the link to the sheet, I moved that to the top of the post and included a screenshot of the current version which has changed a lot since the original post.

The comparisons to the old Storj version are no longer relevant, but just for historic reference I kept the original post at the bottom of the top post. Hopefully these changes will help people find what they need without being distracted by no longer relevant comparisons.


Should caution that the 2.5TB default for “Max Ingress TB / month (estimated network activity)” is still very optimistic since it is assuming that test data is being pumped to SNOs every day of the month. This has not been the case for the last couple months.

It doesn’t assume that. My nodes have received 2.54TB last month. The last few weeks of that month saw much reduced ingress. Month before that over 2.3TB. It’s based on a recent average, not constant ingress.

Just saying, it is still optimistic. One of my nodes is going on 4 months old and it is no where near the $16.64 total earned (3 months total earnings).

One of your nodes? You do realize nodes on the same IP share traffic right? Sum them up and it should match.

The node I linked is just my one node in a separate location from my other nodes. But I have no way to tell if someone else has a node in the same subnet.

That is weirdly low ingress. What kind of connection is this node on? Did it have free space throughout all months? What hardware is it on and how is the storage connected? And if the node has been updated to 1.9.5 yet, what are the success rates like?

Also, when was it fully vetted?

It is on fiber 1000/1000. Only 2tb allocated but as of today, only 1.87tb full. So I do not believe it ever hit the max cap. Yes, 1.9.5 as of today with linux release. And this is a separate node from my trashy one.


I can’t explain the difference then. Maybe other nodes in the subnet. How does it compare to your other nodes? Are they seeing closer to 2.5TB ingress in July (per subnet)?