For the last half a year there has been a steady decrease in stored data.
The amount of data on all my nodes as of today is back to August 2022… In the last 6 months, more data has been deleted than added. Where’s the promised network expansion? There’s some kind of degradation…
2 Likes
1 Like
Is just Trump getting rid of population surplus… 
That was January 2024 now March 2025. The rapid decrease in data has been occurring over the last few months
1 Like
I have had over 10 TB deleted. And the deletion is ongoing.
I have been seeing the same, nodes are pretty much stagnant in stored space.
I recently build a version of my website (this part is not public at the moment) to monitor this exact thing.
I will log all storage stats (daily ingress, egress, payments, total held and stuff) daily. I will do this every day, to get a documented history of how my nodes are doing.
Seeing it on a graph, when I have many nodes til be useful.
And that way I can see if my total storage stays, decreases or increases.
Hopefully someday I will make this part public
the network is dying and they only focus on their Select network. Lol. No data for the public network.
5 Likes
Anyone else remember when they wanted us to book Western Digital’s production lines exclusively for SNOs because we would not be able to add storage fast enough? Or that one time that they asked us to all have 100Gbps dedicated internet lines (multiple), directly wired to our ISP’s core routers, because the SNOs would not handle the incoming traffic and they were worried that the mythical client would not choose us?
The network is dying because of two reasons: they simply refuse to stop new nodes being created (my simple estimation shows that there are currently 15,000 more nodes than needed) and their laser focus on the wrong type of clients.
Storj is perfect for one single (and simple task): long term storage. Any other use case is not for this network.
None of us expects the network to grow in the foreseeable future (=the actual definition of foreseeable future, not the 2-month-Storj-announcement-definition).
3 Likes
I think just looking at the last six months (after the test data)… if you try to ignore all the ups and downs with bloom filters… the average node is still growing about 1TB/year. And we’re gaining around 200 nodes per month (averaged over that year)?
So, not nothing.
(Edit: With node selection now having some performance tracking included: perhaps SNOs that don’t see growth at all… are just seeing data naturally rebalance away from their (presumably slower) nodes?)
3 Likes
There is no growth really. I don’t know how you guys are seeing the (non existent) growth, but here i the truth from actual monitoring systems that aren’t based on artificial random dashboard metrics:
The dip down to 20% can be ignored, it was the disk being unmounted for a filesystem check.
I can take screenshots of all of my nodes, but it’s basically the same thing. No new data, data is on a clear downwards trajectory. You can extrapolate the graph out and see when it will hit zero (it will, eventually).
3 Likes
It would be possible for ordinary users to store their data if they had a native storj application, but there isn’t one, so let’s settle down.
The opposite is true. Storj is perfect for handling short term geographically accessible large blobs of data. Why would anyone use hot storage for long term storage?
I think you are conflating “what SNO want usage to be” with “what customers want and works well”.
For long term storage there are much more cost effective services
Yea, why would wikipedia want to store data on hot storage? You can always sift through your old encyclopedias to find what you need.
first and second year than numbers changes…
Well there is also the fact that recent improvements made it so that even my Pi5 has an awesome success rate. So maybe your fast machine had an advantage and the improvements reduced that advantage. Just as an alternative.
1 Like
Before continuing with wrong suppositions, let me clarify the situation with real data, from tracking my farm since 2021.01. I currently run 18 nodes in 10 /24 subnets.
The number of nodes and subnets are pretty constant in the last 2 years.
So stopping new nodes from onboarding isn’t a solution. They are not to blame.
The stats are caused by:
- free tier data removal;
- test sats and data removal;
- payed data constantly increasing since day 1.
The number of paying customers is somehow irrelevant and we don’t have those stats.
Only their data is. My current payout is double the electricity cost.
The first 2 charts are for the entire network, the last 2 are for my farm.
1 Like
What I’m seeing is that data on EU1 and AP1 are going down. The only satellite that is showing data increasing is the US1 satellite. What is actually happening is that US1’s select network is roughly matching the overall deletions. Since we mortal SNOs aren’t part of select, what we see is data going down and down and down, for the past 3 years. A small pool of data, being distributed with an ever growing pool of nodes (yes it’s only 2K nodes up in the past 2 years, but up 10K nodes in the past 3 years) means that each node gets a smaller piece of the data-pie.
I peaked at 4 times the income 3 years ago. Can anyone offer any insight as to why my payout is 4 times lower than what it was? Yes, the answer is data is shrinking.
Until Dec 2023 (not even a year-and-a-half ago, yet) Storj payout rates were above what they were earning from customers: and SNOs were bleeding out all the treasury coins. The farther back you go… the more nodes were being overpaid.
High historical payouts… were never because of high historical usage. It was “free money”: but those piles of tokens were burned through by now…
Edit: This data has a lot of dropouts: but it’s the node “Used” number from the last year.
Ignoring the massive spike in Test Data… doesn’t it still look like data is increasing?
2 Likes
i don’t see much data shrink.
Maybe last 2 months, because the long awaited deletions of old free accounts, as @snorkel mentioned.
Grafana currently shows Network at 1/4 full. (because we added more and more free space)
Even if there will be new data up to 3/4, the network must have been set up for a node to never go full IF the network has collectively a space left.
Because over all my nodes with 4TB-16TB disks, all Nodes are around HALF full, and disks are storj dedicated. And this is over at least 1 year back.
(Some exceptions like one is 1/3 full, and others 2/4, one closer to 3/4)
it means to me, that ALL data is spread over the network pretty even (to prevent a node that is decent (which almost all are in normal situation) from clog, to always has free space to welcome potentially new customers.)
So if there will be NEW data, like say someone comes and use +20% of current free space nodes got, it will trigger eventually creation of new nodes (a signal to grow), resulting over time that temporal spike, to be still approximately flatten for everybody.
Therefore i assume: In current situation, there is no way to get Your node full.
It wont just grow past some % of total network free space.
Bad news for anyone hoping to get $ from occupied full disk space.
Therefore the only hope is in increase of Egress (paid 2$/TB)
BUT thats pretty well guarded by current End price at 7$/TB for customers.
As a SNO i would like it to be as close to 2$/TB which i’am getting, to lure more of the market pie.
But as a SNO i don’t have the full picture, and the recognition of the market situation.
I know big tech collect at much more rates!
BUT it might be completely not the matter of price, for “why not everyone uses our services already!? its sooo good”. (Maybe its just pure politics? and a matter of time?)
Either the statement that 7$/TB egress is cheap compared to the market realities is true, and the usage will march forward, just matter of time.
or not. In both scenarios we have no control over it. Because its the management, who have full picture and to make decisions and run the company. And we are just offsetting our already running machines, voluntary joining the storj network. We just enjoy whatever it is. Aren’t We? @Mitsos
So please, pleaseee, don’t write any dramatic posts, “oh how it shrinks” or whatever. Don’t make a false impression, that the network is not used or somethings wrong. The network is used all the time, and as a small SNO, i see more and more egress form last 4 months at least. Because later some potential customers reads that and gets wrong impressions. Its just counter productive.
If You a SNO, just let the ppl work.
There are a lot of improvements over just 1 last year!
Look at the amount of updates from release to release.
can’t wait what future will bring for the team.
They work hard.
And we’re here rather to support the team.
If You have propositions for the better, then please post.
6 Likes
satellite reported (before expansion):
EU1
-10PB in the past year. Clear data shrinkage, not a single person on Earth can deny this.
US1
Wow, business is booming! +8PB in the past year. You are absolutely right, even I can’t deny this shows clear data growth! I stand corrected.
AP1
Not bad, down almost 50% but climbing. Overall let’s round it up to -1PB.
Let’s see: -10-1+8=-3PB.
Saltlake doesn’t count, obviously, but here’s that as well just for completeness of reporting:
Now scroll up the thread a bit, where I said that the only network growth is due to US1’s select network growth.
2 Likes