I’m not so sure about that. Do you realize that there are 32000 nodes now (many are fake), but 82000 have quit or been disqualified over the years? It seems a shame to have lost so many nodes. We could debate the reasons, but it seems like there is, or was, quite an interest in a project like Storj. Maybe if cheaters weren’t bleeding smaller, less sophisticated nodes, more “use what you have” nodes would have stuck with the project.
I agree that it’s hard to attract corporate interest in a project like Storj. A suggestion I’d make to a new owner is that the forums be separated into Customer and SNO/developer forums, with restricted registration to the latter. Customers don’t need to know all the drama that goes on behind the scenes, or see the pictures of a SNO running a hard drive off their Roku box, as cool as that may be for techies, hobbyist, and hackers. That image won’t impress a corporate CTO; the reliability and global distribution will.
The reason are very obsvous and simple. Initial bootstrap with unsustainable payout rates attracted wrong type of operators. They were appropriate at that time but undesirable as project matured. Nodes are designed to be replaceable and disposable. It works as designed.
It’s not hard. You just need to have a working marketing team. The reliability (time proven - -they haven’t lost a byte since inception over a decade ago) and performance (no competition effecitly for some use cases) of storj speak for themselves; it’s hard to mismanage, but as we saw, not impossible.
The current situation is 100% marketing failure. Engineering is top notch.
@alpharabbit and @arrogantrabbit this is a bit off-topic: but I think you guys have monitoring. Have you seen an OMGWTFBBQ-level of egress over the last hour?
Removing a filtering rule would create an incentive to create more nodes, because it will naturally increase the percentage of the probability to take a bigger share from representation. However, choice-of-n rule might help with that a little bit - slow nodes will be selected less often. So placing multiple nodes to one HDD will have a natural limit, especially during compaction and attempts to handle multiple customers’ requests - their success scores will be low.
Also, the current big clusters will get most of the uploads due to their share in these 30k nodes.
I would prefer, 100 corelated professionals, than 10 000 uncorrelated who dont know how, or dont care. when you make some $ you really dont care. when you start making few hundreds or thousands- it start be more interesting.
when you understand it, and want it long term- then you understand that it should be profitable for all parties also.
we live in real world, not in ilusion.
with today prices for electricity- no one hold online capacity not used.
So lets stop treat it like hobby party and start it like normal business model.
Storj needs most cheapest supply- but today 20%-30% of all storj payouts go to VPS providers.
problem is /24 and even /32 will do the same but in lower proposion(because people sooner or leter start to use own ISP IPs as VPS go more expencive.
we can remove IP filtering
but there is problem correlation of nodes will rise uncontrollably- this is bad need mesures
there is lot of positive sides like network will work faster, can lower prices.
Geofeaturing will start really working.
Storj will start see corelated nodes, and can make some repair to mitigate that.
The network is designed to operate in a Byzantine environment. It’s better to have fewer correlations between nodes than more concentration.
We don’t treat anything as a hobby; we simply don’t recommend investing in hardware solely for Storj; that’s the difference.
And the price reduction revealed the exact reason: the owners of large shares of the data are trying to dictate the terms. They’re smart, have technical knowledge, and perhaps better equipment and setups, but if they start controlling a large portion of the network, it’s bad for the network. We need to ensure that when they shut down their nodes, it doesn’t harm the network.
Therefore, I still believe the /24 filter should remain. And something on top should be added to distribute the data more evenly, even if there are Byzantine/rational nodes that have managed to bypass the filter.
We’re not against their presence; we just don’t want them to concentrate correlated data.
If there’s nothing worth implementing for now, then let’s raise the filter to /23, /22, …; This will result in significant redistribution of data even when using only one step in the filter.
I hope Storj know it got that big due to lots of people treating it like a hobby and buying hardware just for Storj. I don’t think it would’ve gone that big, with just random 1-2TB nodes hosted on Synology or QNAP NASes.
We always said - don’t buy anything exclusively for Storj. Did you ever saw anything else from me?
It might be profitable, but it’s not guaranteed. Likely - not profitable, if the only goal of the setup is to get profit from running nodes.
@gingerbread233 if there is a way to bypass anything to get more profit - it will be done. Until it’s not profitable. This is why we got all these discussions about electricity costs, hardware costs, etc. Because there is a lot of people, who didn’t listen to us, they decided themselves, but for some reason is blaming us.
Yes, now Jammerdan will come with their “we need more customers”, that’s right, we need more demand than supply right now. And we need to cut costs, so thus this topic.
I agree, we need more customers, they come every day, some stay, some don’t upgrade, however, customers are growing, but slower than we want.
I don’t blame anyone in anything. what i buy is my responsibilities.
I try to think how to make it more profitable for me, for you and for storj.
If operators can reduce costs, storj can reduce prices.
I just try to think and find solutions out of the frame, that you try to stay in. i understand It looks safe and dont do anything. but business dont work like that.
If you read upper posts, I try to understand also the probability that will be problems with this decisions. But without exact info how storj select nodes it not possible.
I understand that @littleskunk have lot of work and dont have lot of time to explain us all details of this algorithm. May be at some point he will find this time(i hope) then we can also calculate possible problems and think how to mitigate them, but sitting in the frame like storj do last 5 years bringed us to chapter 11 today.
but I also understand, that not losing client data is essential.
but efficiency is key feature of every system, without it nothing can survive. today storj is not efficient. if 20-30% of all cost go to VPS providers
I believe that not the selection algorithm or any other technical issue bring us here. Here I agree with most of suggestions from the forum members.
The current topic is just reduce costs, if it also would help to bring more new customers - it will be wonderful.
I hope that the proposal to bring more customers will pass a review soon and will be published. I think that it might help.
That is not what I am talking about. I am talking about the amount of data a single node is receiving. Just put it this way: If a single node would receive so much data that one disk would be constantly saturated, then nobody would put more than 1 onto 1 single disk.
The promise was: SNOs provide disk space and Storj takes care of filling the disks. And Storj did not deliver.
Instead of filling our disks, we are discussing how we can distribute the miserable 50+ PB among the operators. And of course everybody has an “idea” that suits his setup best. No wonder, the community is destroyed.
Now the /24 rule is a setup decision. The rule shall prevent loss of customer data. That’s all. Now data proves that data loss is not to be expected without it. So my opinion is: Let it go. Now we can treat any node as independent from any other node and don’t need an artificial limit that proved it did not work. Neither does it help to make customer data safe nor does it help against cheating.
Maybe dropping /24 even helps with that when nerby IPs are as good as far away /24 IPs.
Also the tech will give the customer the fastest nodes anyway. It has been said in the past so many times that VPNs make things slower or don’t work at all. Additionally multiple nodes on a disk shall be slow down all nodes on a disk. So many nodes on a disk routed to obscure VPNs should not be that great as result.
Therefore of course, those who have their /24 ips already in place don’t want that to be changed as they’d lose their advantage. Drop /24 pass that advantage to anyone.
But at the end it is not important as the network takes care of it for the benefit of the customer.
You can’t.
And just to put this into perspective:
Storjs Chapter 11 attorneys received $31,950.00 for fees incurred prior to the filing of
the Chapter 11 Bankruptcy.
Storjs Chapter 11 attorney just asked for compensation of $21,181.50 for their work after July 31st until today.
Accountant Heather Trowbridge asking for a fixed fee of $39,000 for June 2026 - August 2026
And Dr. Ali expecting a compensation of $10,000 per month + a percentage from the sale of the company.
After destroying customer trust with that $50 minimum move now they are destryoing SNO trust and community. At least that they are doing very well.
Let’s hope somebody is migrating away. At the current stage of the company, any customer who does not save their data somewhere else can’t be helped.
If heated up very slowly the frog will not jump out.
That is the frog suddenly dying.
Which helped exactly 0 to prevent that current situation, right?
That is a common problem with companys that are run be tech people.
It probably would be the same size like Sia today.
Of course I will. Because this is the reason why we are where we are and it is the only way out.
And I keep repeating it as the company is not making a single step into the right direction. They did not in the past when they had enough money and they are not doing it today.
And I keep repeating it because it is an strategic decision and a path that must be actively worked on and an approach to simply wait for customers will not be sufficient.
Unfortunately the company still does not take that strategic approach.
Instead they fired their sales team that they had just build or increasde at the end of 2025.
Now the company is run by 1 or 2 devs and suddenly we are told that this is totall enough to run it, to keep the current customers happy and it would be even sufficient to do some development.
So the correct strategic move would have been always: A maximum of sales people + a minimum of tech people. At least the impression we received from the company was completely the opposite: Tech people all over the place but zero sales people. And the result is what we have today.
If this does not get resolved the company will be back in Chapter 11 as soon as it emerges from it.
