5 USD per TB egress would be a too harsh cut from my view. There could also be some kind of a leveled model, with multiple earnings in decreasing. I would prefer this over a single cut.
Apart from the cutting down the revenues there should also two things happen:
Reduce/lower the free offer of STORJ DCS. From my view all data and traffic can be paid enterprise data and traffic. I don’t think free users help STORJ DCS grow to exabyte scale. So the current offer for free users is kinda too high and should also be lowered.
Remove the test traffic. It is rumored that saltlake and europe-north are test satellites. Why cant the data slowly be removed and after full deletion turned off and gracefully exited for the two satellites? For some reason the gains from these two sattelites are kinda good on my old node from 2020-10:
Last quarter STORJ tokens are paid out 9.8M from the reserves (non-circulating supply). At this rate the reserves, assuming they can all be spent, will deplete in 20.8 quarters, or 5.2 years. Now I’m not saying STORJ is a tech start-up. But if it were one, 5.2 years is like two lifetimes… and a half. So I’m not worried.
Loss making is not preferable but unfortunately that’s just part of life, especially when you are a wholesale low-priced new comer in saturated cloud-storage markets. Backblaze, for example, has net profit margin of -35% (Q2 2022). It is making loss, can’t pay debt, and survives on new capital injection from shareholders… for now.
Increasing prices too soon, or too far, you risk loosing traction, customers, and the network effect - everything a new comer, like STORJ, desperately needs. Reducing payouts for node operators, you risk loosing supply, community, decentralization, and stability of the system - everything STORJ differentiates itself from others.
Don’t do those things unless it is absolutely necessary. Now it is not. At this early stage of STORJ, it is not about making profit, not even about breaking even. It is about hanging on, waiting for that iPhone moment.
So STORJ is in kind of “innovate or die” situation just because everybody is. It is in a better position than most in my opinion. The hope is that something good will happen like:
the developers come up with a new tech / product / technique that can either lower cost or increase revenue
competitors go out of business and then STORJ can raise the price
STORJ token prices go up —> less STORJ paid out for the same amount of USD
Given the lifeline of 5.2 years, each of the above is a real possibility. So, again, I’m not worried.
The test satellites are there for a few reasons (other than just testing), but mainly to provide more income to node operators and to pre-allocate some space. If a large customer decides to upload 1PB of data quickly, then data from test satellites could be deleted and the space used for that customer. Otherwise there is no real way to know how much space is there on the network (I have not allocated the entire pool to Stroj, I increase the virtual disk by a couple of TB once in a while when the node runs low on space).
I think for node operators Storage can be 4$ per TB and then egress can be 1-3 $ per terabyte.
Then operators can have stable income and Storj can make money on Egress.
The energy cost is a good think for us.
A big company has no incentive, they just rise the cost for the end users. Recently they did exactly this and others will follow using the “all did this, we will rise cost too”. If the competition, normally, makes the cost smaller, now it all doing the opposite…is a race to increase it ^^
The nodes are not big companies…they are individuals, and as such, they are faster, smarter, and have all interests to maximize the profit. To make an example, is the same thing that happens to bitcoin networks years ago, to maximize the operation, the majority moved to renewable energy.
Even on a small balcony, you can put a 200W and a battery and power your nodes for more than 10h…I would challenge a data center to do the same, it would probably run out of space powering nothing.
In general, we can’t keep fighting at their rules…we must rewrite them, and challenge them where they are more vulnerable.
The vetting process
We already spoke about this a lot of times, the current system is broken at the best.
I have some nodes with 300$ held each one and others with 20$…all fully vetted. Maybe is time to rework this system eh?
A possible better model is to force the node to keep at least an amount equal to the payout and until this, pay just 50%…without 15 months or any other time limit.
For example:
Month1, make 20$…10$ got paid and 10$ get into escrow.
Month2, make 18$…9$ got paid and 9$ get in escrow (19 total escrow).
Month3, make 19$…the node has 19$ in escrow, so they will gain 100%…so all 19$ will get paid.
Month4, make 20$…the node has 19$ in escrow, so they will gain again 50%…so 10$ will get paid and the new total escrow will be 29$.
Month5, make 22$…the node has 29$ in escrow, so they will gain 100%.
So, the escrow follows the profitability of the node…this has more sense and should decrease a little the repair cost.
If not this model, at least a variant of this…
Price model
The current price model is fair, the only problem is that this works only for offline storage (mostly backup). But is prohibitive for a working copy (like dropbox does). Maybe a solution is not to decrease the price per TB, but to create multiple models and node types and let’s operator choose what is best for them. For example:
A 250GB SSD node, can opt to get paid a fixed amount per month divided by the total number of nodes (one per IP address) (like 25.000$ \ 10.000 total nodes) BUT receive no bandwidth payout. A node like this, acts like a cache node that downloads most required files from other nodes and let’s end user to download\upload them without worrying about bandwidth usage (always using erasure code and multiple nodes to ensure durability). Such nodes are responsible mainly for reliability and will move files to storage nodes as soon as possible, also are not mandatory for the networks to work.
Other nodes, can prefer to have big storages but get paid per GB and bandwidth. Because maybe they don’t want their network to get spammed with too many requests. Also because the equipment has a cost. These nodes are responsible for maintaining network safe. The egress is still paid, but usage should decrease significantly. Also, these nodes can replace cache nodes if networks become busy (but egress must be paid).
Others can offer both types of nodes.
In such way storj can turn from an offline/backup system only, to also an online persistent backup.
For business backup, i think storj is currently the best offer for both individual and reseller service (like all integration that currently storj supports).
But is currently unusable for the rest, maybe we can boost the network usage if we support different kinds of nodes\price models.
Hard to say to be honest. It is most likely what has sparked this much interest. And the prices aren’t that much lower than price competitors like Backblaze B2. I understand why they did it. It sucks that cost of electricity has gone up for a lot of SNO’s though. The timing is bad all over the place.
There’s not much margin if they want to stay below cost competitors. Backblaze B2 is $5 for storage and $10 for egress. They would have to stay below that. And that’s ignoring the bad optics of raising prices now.
Me too. But I figured this would almost certainly not be the case for many others. So I wanted to get that discussion started.
True, but they could announce a period of surge payouts and while you can’t scale running costs on that, you can still use that to invest in expansion of HDD space.
It’s gonna be hard to drive up demand enough for that to matter. Most tokens aren’t spent on paying node operators atm. So even if every customer paid in Storj, the circulating supply would still grow, not shrink.
You sure have a way with words… You definitely said something here, but I still don’t know what this means in practical sense. This doesn’t even confirm whether payouts will go down at all. But I get it, this probably needs to be discussed internally before something can be said publicly. We’ll just speculate in the mean time. I hope that’s still helpful.
That may be true, but it’s still a limited resource. Not charging customers for egress would lead to a lot of CDN like usage. Which could bottleneck nodes hosting popular files. There is a balance to be struck here. @Toyoo also made a good point that egress payouts incentivize node operators to provide fast egress.
I agree, but egress has already gone up quite a bit. At any moment something like this is introduced it will still feel like a big drop. That’s why I suggested easing into the new payout scheme with the help of surge payouts.
I like that idea a lot. That would definitely help soften the blow and prevent existing node operators from leaving. (though it’d make changing my earnings calculator a bitch, lol)
I don’t think this is ever a good idea. You’re essentially putting a hurdle on adding more data to the network… meaning a hurdle that would prevent future income. It would also be a cost that no competitor charges.
I wouldn’t go that far. It’s a tough puzzle to crack, but I’m sure there’s a good solution to be found. Node economics look really good now if you run on hardware that is already online. You can earn back a 4TB HDD in about a year. Even with earnings cut in half, you’d get there in about 18 months. That’s not bad. But yeah, they have to ease into everything to prevent churn.
Keep in mind they only get paid 7 USD by customers. Anything much higher would not be profitable. But yes, I agree they should slowly drop to that level. This is why I suggested compensating with surge pricing during the first year.
Not a rumor. Only eu1, us1 and ap1 are customer satellites. The others are for testing/incentive. Removing test traffic would reduce costs that Storj isn’t getting paid for by customers, but doesn’t resolve the fact that they are also taking a loss on customer data/traffic currently.
They mimic customer egress levels intentionally to make the data not less interesting for node operators (for the most part).
Because no new data has flown into these sats for a long time. You basically just get data from them through repair.
It doesn’t help that the number mentioned in the bullet points also seems wrong.
This may have been an expensive month though.
I would say no. This makes sense. It’s just that the timeline is shorted than what you used for your calculations. However, we’re just talking about token supply here. Storj Labs probably has other sources of funding as well. So who knows, they may be able to afford taking a loss longer than we think.
Total earnings per TB average around $3.30 atm… if you include egress. Your suggestion would have Storj take an even bigger loss. It’s also a bit of a risky move as it really relies on Storj getting very specific egress heavy customers to be able to make a profit.
Absolutely! Anything to reduce unnecessary repair costs. I think they should not match payout though, but rather repair egress costs. So you would hold back 50% until the amount reaches $5 (assuming prices I mentioned in the top post) multiplied by the number of TB stored. Double it for extra incentive.
It’s really hard to verify remotely that a node is what it says it is though. I have an SSD accelerated array. I could claim it’s all SSD and many performance tests might even confirm that. I think you would have to differentiate on measured performance, rather than a class type split.
Keep in mind that over time, the amount of data stored on my node has increased, so mbps/TB probably went down (though this depends on the usage pattern, maybe nobody accesses the old files, only new ones).
I guess nodes could specify the class (storage, cache) independent of the actual hardware and then the slow ones will just “lose the race” all the time.
If it makes sense (panel+inverter cost vs electricity cost) for 200W, it makes sense for 20kW.
That number is wrong indeed. I use the old number from Q2. Apparently Q3 is much worse. The time reduces to 2.6 years. Now that’s interesting. It’s not imminent but not too comfortable.
Still, the good things I listed can happen even with the reduced time. Notably, the price of the token is very depressed right now and likely already forming a bottom in the midst of a brutal bear market. The STORJ tokens paid out from the reserves are thus higher than usual. Once it recovers, hopefully in a few months, STORJ can have more time.
BTW, if it comes to make a choice between adjusting prices/payouts and raising capital from external funding, I would go with the latter. Issuing new tokens, if that is possible.
As SNO I’m open to support the project in the hardships, as right now Storj will be slowly pressured in his current position of supporting SNOs and clients.
I have a raw idea too, feel free to ignore it. lol.
SNOs that operates during hardships (bear market, lowered prices…) get a plus in reputation somehow rewarding them when great times arrive again. Hopefuly this is not even necessary.
Thats sort of done unintentionally right now since we’re paid USD amounts in storj. If you hold the earnings during a bad time (now) to a better time when the token price is higher, you got paid more for being there in a bad time.
I started my first node 2 years ago. It started quite slow, but this year it gained momentum and I just started my 4th node (behind same IP) as I was about to run out of space.
In October I had TTL 850GB normal egress and 137GB repair egress with TTL 8,4TB stored data. I will receive around 26USD for this. It was about 2 months ago when I started to feel that I do for some real monthly revenue instead of doing it for fun (and supportive curiosity).
I believe my free time what I need to invest (and I spend) into STORJ should be rewarded.
I know the evergreen saying about leftover hardware, but node operator’s time should be considered and paid… I don’t mind the electricity cost and the whole house is running on solar and I have zero bills (yes, I know about the investment and return, but let’s put it aside now).
If I would receive half of this monthly payment, I would quit quite quickly.
I don’t know what is the avg node age, stored data and monthly paid amount, but to do it for 5-10 USD/month… it simply does not worth all the hassle
I also see that as soon as the node is full, the revenue drops to about 50% as old data is rarely read.
Frankly, I don’t understand the fear about that the network would run out of free space. According to STORJ Network Statistic, 13.8PB is used and 23.2 PB is free space.
Yes, I understand the node operator incentive for the stored test data, however still, it would be better to delete these to have a more clear picture.
At the end of the day, all of STORJ representatives’ comments are quite misty, so if they are happy with the current financial situation, I’m happy as well.
As soon as they present a REAL business plan for the node operators, I’m ready to comment it, either by staying or leaving.
I dont think that storj need to compeet with backblaze, they cand make this speed that have storj all over the world. And there is no point to sell amazon speed solution for the price of slow backup solution.
Also i think there is possible that people can choose expansion factor, like if you want cheap then expansion factor only 1.5x , if you want big redundency then full expansion factor. then price is other. then you can aim different tier of clients. Bigger expansion factor also mean faster download back to client. as there is more pieses in network.
It’s not so much a situation where we’re going to announce reducing what everyone gets paid, or whatever. That may be part of it. But there are a number of initiatives here that are being worked out. What sticks and what doesn’t is part of those strategy sessions that are ongoing and will then be brought to light on Twitter, Blogs, and here in the forum of course.
Consider that we may need to have a more professional type of SNO. One that signs legal agreements and certifies their hardware. Larger operators that can partner as both node and satellite owners. And a system that could “spin up” or “spin down” storage on demand. Certain customers may pay more for geofencing and to have their data on the certified side of the network. They may have needs for high demand times or locations, and that could have different costs and potentially increase revenue for SNO’s.
And also, free tier is probably also unsustainable for the long haul. Perhaps limiting how long something is free before charging or not. I expect changes there.
Of course, for such things we’d need better tooling and node changes to make multinode easier to manage. And Satellite changes. And all of those kind of things need to be hammered out.
And consider, if Satellite operators can make money without using loss leader tools like the token. We can find more Satellite operators being spun up, more data being added, and capacity increasing significantly.
So, I would probably wait for more info from the formal announcements so everyone can have a better idea of the direction we’re going in here. There is a lot more to it than just changing the fundamentals of the existing pricing. As John mentioned, we want Exabyte storage and we’re not going to get there if nodes can’t earn. Just how we are doing it today is going to be changed around, and I would expect there to be incentives for different aspects as it rolls out and permanent ones as well. There is a lot of topics around this in discussion, not all of it is going to happen, and not right away for others that will.
At this point, it’s best to expect that change will happen, but we’ll need more info before really understanding what it is and how it’ll impact SNO’s. But the plan isn’t to pull the rug out from under you. SNO’s are important and the life blood of Storj. We want everyone to earn, and that includes SNO’s. Let’s wait and see what comes, a lot of it is quite exciting sounding and I think people will be happy with a lot of what is coming.
Storj team you like to tell anything but at same time nothing specific. Same as on town hall, just mentioning some changes but nothing specific, then people always will think the worse scenario.
If you push out jeany from the buttle I wish to hear the detail’s, for example about professional SNO’s what will be hardware requirements for this and other deteils.
I get that you can’t announce anything yet. But the way it was now mentioned without any context whatsoever, yet actively asking for feedback is just a little weird. I’m with @vadim on this. We are going to think the worst.
For example, you mention professional grade storage nodes and my first reaction is to think that us reliable but not professional node operators will miss out on a big piece of the pie or even be pushed out. Now I know that’s most likely not the direction you want to go in, but without more context, it crosses my mind and makes me wonder.
You can trickle out announcements and teasers when what’s being announced is something positive and new. But with the cloud hanging over our heads of knowing Storj currently loses money on data stored and transferred it’s reasonable to have some concern about the future of payouts. In that scenario, it’s better to be clear about an intended direction. Even if it doesn’t yet include exact details.
At the moment, they do. Backblaze is a familiar name in the industry. Storj is still little known outside of the crypto world (and even inside). For people to bet on an unknown, it has to be better on all fronts. And money talks. When Storj has a more recognizable name and is better known as a good option, this may change. But for now they need to do better at a lower price to break through the innertia.
You definitely said something here, but I still don’t know what this means in practical sense. This doesn’t even confirm whether payouts will go down at all. But I get it, this probably needs to be discussed internally before something can be said publicly. We’ll just speculate in the mean time. I hope that’s still helpful.
