I have thought of something like that too. The problem with it is, that if it’s true, that the gateway makes 95% of traffic, then it is like complete adoption of the new proposal. The remaining 5% are basically nothing.
And it does not change the general failure in the economics of current operation. If payout for uplink traffic remains the same, Storj still receives $4/$7 from their customers while paying $1,5/$20 to their SNOs.
So if there are different schemes, still it would require a significant slash and adds lots of additional overhead and again uncertainty to the SNOs.
True, but not everyone knows that. Especially in companies where to many people are doing one persons job… which is like… most big companies. It just makes it sound a little more complicated to anyone who doesn’t know better. The fact that it’s not really difficult though is a huge plus!
Also a huge plus, however many companies have moved to cloud services in order to get away from managing their own equipment. Although a big enough savings would incentivize this.
Also true… but most of them still have fair use policies. And if they don’t, your connection is usually limited to 10, 25, 50, or maybe 100mbps which isn’t nearly enough.
Good… so we don’t have to worry about people hosting PB’s of data and downgrading to a dial up connection. Got it! To take that a step further though, in case someone still did try it, nodes can just be disqualified for consistently loosing races beyond a certain threshold over a set amount of time.
2 Likes
Here’s a counter proposal.
Keep current pricing model for existing customers, they’re grandfathered in.
Ditch your current charity tier and turn it into a 1 - 3 month free trial.
Lower expansion factor from 2.8 to 2.5.
Create 2 tiers:
Tire 1 (Low BW) - $4/TB & $12 egress. SNO’s get $1.50/TB + egress. Storj get’s $.25/TB.
Tier 2 (High BW) - $10/TB & $2 egress. SNO’s get $3.50/TB + egress. Storj get’s $1.25/TB.
Use cases based on 1 TB of data:
Low bandwidth - 0% to 60% egress per month:
Storj T1: $4.00 to $11.20 Best option.
Storj T2: $10.00 to $11.20
AWS S3: $23.00 to $71.00
Customers save 82.6% to 84.2% compared to AWS S3.
SNOs earn $2.70 to $8.70/TB. (Low end based on 10% egress.)
High bandwidth - 60% to 1000%(+) egress per month:
Storj T1: $11.20 to $124(+)
Storj T2: $11.20 to $30(+) Best option.
AWS S3: $71.00 to $823(+)
Customers save 84.2% to 96.3%(+) compared to AWS S3.
SNOs earn $4.70 to $23.50(+)/TB.
In order to earn even the lowest rate per TB at $2.70 using the high end of what’s been proposed, current egress would have to increase by a factor of 3.4.
Anyone choosing to use Storj’s S3 gateway should of course have to cover those costs as well essentially paying for the convenience factor, however the savings should be enough incentive for customers to run their own gateway.
Now I assume the two tiers would have to operate on separate satellites, so I know the first thing people are going to say is that everyone will just exit the tier 1 sat. Well guess what… right now you wouldn’t make anything anyway. Worst case though, Storj simply makes it so you can’t exit the tier 1 sats without exiting both. Problem solved, go away.
So…
This prevents current customers from leaving.
It pays SNOs decent.
It eliminates Storj subsidizing SNOs.
It starts generating income for Storj.
It still doesn’t discourage lower bandwidth customers.
But most of all it further creates a HUGE incentive for really high bandwidth customers.
Let’s take software distribution for example… how about the gaming industry.
Well use the game ARK which I know is well over 100 GB, but well just call it 100 GB.
They sold 5.5 million copies. So it was downloaded at least 5.5 million times.
That’s over 550 PB worth of bandwidth for a single 100 GB piece of software.
As @IsThisOn pointed out, this type of distribution is typically done using CDN cache. Well… the lowest price I’ve found between Amazon, Google and Azure (links a couple posts down) for CDN cache egress bandwith is $0.02/GB. Tier 2 pricing puts egress at $0.002/GB which is 90% less. And since Storj is already globally distributed, I suppose certain customers could even ditch their CDN cache. Hey, another selling point! Now I assume some amount of data must be cached automatically by service providers or something to some extent, but when you have to pay this much to these companies specifically for this type of service I can’t imagine it’s done for free, at least at that level. I don’t know that much about such things so maybe I’m wrong here?
Now for Storj:
Just for arguments sake, if all current customer data was on the tier 1 plan Storj would no longer have to subsidize SNOs and would already be bringing in $57,300/yr from stored data. Not too impressive yet, but 19.1 PB isn’t that impressive either.
Once they hit their first exebyte though (they’re already almost 2% of the way there), they would earn a minimum of $3,145,728/yr just from stored data if all the data was only on tier 1. That same exebyte on tier 2 would make them $15,728,640/yr. And none of that even accounts for other forms of income they may have.
Your not going to sell customers on Storj by promoting end to end encryption, “great SLA’s” or the fact that the data is nuclear war proof, and the crazy low prices are just going to raise questions as to how and why? And let’s face it, telling a bunch of business professionals it’s because their data is hosted on a bunch of morons gaming machines isn’t the best approach to this. It’s inexpensive because this unique way of distributing data in a secure and encrypted manner opens up a ton of otherwise idle bandwidth to Storj from all over the world, which just so happens to have the added benefit of being both highly available as well as highly immune to service interruptions not achievable by ANY of todays datacenters… would be a much better answer.
Storj has one huge selling point over it’s competition… lots of otherwise idle bandwidth that allows for dirt cheap egress and fast connections. Use it! And stop freaking out about a little price increase, your product is more than worth it.
8 Likes
Oh shit, I forgot this is racist now, lol. To bad.
Took me a while to read all those messages that piled up during the 4 days.
TLDR:
1.5$ Tbm
3-5$ Egress + Repair
Location: Germany
Electricity Price: ~0.33€
I personally could live with the slashing of egress. I already have Storj egress handled with a low priority as to not get in the way of my more important upload. So Storj only uses what is available anyways. A bit of incentive will be needed to not fully block egress traffic but other than that I am okay with reduction here.
Storage costs are something else though. The 1.5$ currently offered were around the amount i felt comfortable using disks to fill with Storj especially due to the added high payout in egress. As said in other comments you cant really easily reduce the size so it’s not really possible to use space you “currently don’t need” as once it’s full of Storj data you gotta keep it for a while. Especially with the slashing of egress these 1.50 now will have to be the backbone of my earnings. So a reduction here would really hurt.
At the current maximum proposed price of 1$ i would have to think strongly about ever increasing my node size and might just continue to keep what i have and just GE each node once i need the space. (I would really appreciate a way to reduce size by just sending my parts into the network and having them be redistributed to other nodes sorta like a GE for a small amount of data without getting any payment for it but also not loosing my node)
I think if Storj really wants to scale it should at least be economically viable to buy storage and make a ROI. I find the mantra of only doing it if you already have a server that’s running 24/7 fine. That way we only have to look at electricity cost and purchase price of the HDDs.
How to do it financially for Storj:
I think Storj needs to increase price per TB by at least 1$ on the customer side to be able to continue to pay the 1.5$. customers already onboarded or in talks can have the old price for now but i think new customers should get a slightly higher price.
Lots of other good ideas where posted but this thread is already very long so I’ll not go into depth about them ^^
Mostly agree with your calculations. For high egress applications that don’t use S3, STORJ could be a cheaper alternative, because traffic is “free” for nodes.
But there is the S3 problem. Hosting your own gateway is not something customers will do! Then you move the costs to the customer and STORJ is only the data backbone. Then you moved so much of the costs to the customer, hosting storage would a very small part of the calculation. If your gateway costs are around 10k a month, you don’t risk loosing your data just to save 100$ by switching the storage backbone from AWS to STORJ.
That is narrowing down customers to a very specific small niche.
No. That is not how this things work. There are cache servers and CDN. You don’t download ARK from one single root server. You download it from cache. Just like the world does not watch Netflix from a server in California. You watch Squid Game with thousands of other people from your ISPs internal cache server that Netflix provided to them for free.
That is why ARK did not pay 550 000 * 20 = 11 Million dollar for bandwidth costs. If the internet would work that way, you would’t get a 200mb WhatsApp update every other day.
1 Like
Why would their gateway cost 10k/mo?
Oh I’m sorry, I didn’t realize all that CDN bandwidth is free.
You don’t download from Storj from a single server ether, so this could essentially eliminate the need for cache in certain cases.
Show me where CDN cache pricing can beat Storj on either of the 2 tiers I proposed.
I’m not an expert in this area so maybe I’m totally missing something here, but I don’t see a price per GB lower than $.02. Now I’m sure it can get a little better with larger packages, but can they compete with Storj?
Cheapest CDN I see: $0.02/GB
Storj T1: $0.012/GB (40% cheaper)
Storj T2: $0.002/GB (90% cheaper)
So I suppose that’s actually another selling point for Storj. For certain use cases theres no longer a need for a CDN cache, and the bandwidth is actually still cheaper on either tier. Again… unless I’m completely missing something.
That’s all HDD responds to when working with STORJ.
At the same time, for working with CHIA, it will be only linear reading and writing.
As a result, the hard drive has hundreds of times more wear with STORJ than with CHIA.
1 Like
These are two very different use cases. Netflix did that in order to get rid of latency and peering bottlenecks. While Storj could help for the latter, it won’t solve the latency issue. However, for downloading games, latency doesn’t matter and it’s all about throughput. If steam had a native integration, it could just use Storj without CDN in order to download games from the network directly and it would be capable of serving most games at high speeds (peak demand for a specific game might still be an issue until dynamic scaling is implemented on the network). So yeah, one of these cases could be a great fit for Storj.
4 Likes
16 months after joining as a node operator, my setup only begin to feel profitable. If the proposed pricing was to be adopted, I would stay as a default bitter decision. But I would leave the instant I could imagine a better use of my hardware.
1 Like
I don’t have to show you anything, you have to show me potential customers for your proposed pricing.
We both agree that bandwidth for nodes is “free” and because of that have to compete on bandwidth. I think one group could be backup users and proposed a price, you think that it is Steam selling ARK and proposed a price.
That does not mean that a single game like ARK creates 1’126’400$ (if we use your proposed 2$ per TB) worth of bandwidth. But yeah, maybe you are right, and ARK really used one of your listed CDN for 0.02$ per GB. That would be 20$ per TB or 11 million $ for ARK. Do you think it is realistic that ARK (or steam) paid 11m $ for bandwidth to deliver ARK?
Latency also does not matter much for video, clients have cache. Sure I don’t wanna wait 20s after I clicked a video, but up 3s is fine by me.
Really makes you wonder why customers choose otherwise, doesn’t it? 
As we have discussed before, customers are just too stupid to realize that they can get a product equal if not better than AWS S3, for 80% cheaper! /s
i phrased that poorly, its the 1.9mil Storj tokens in the other services category in the StorjLabs Q4 2022 token economics report.
vs the 0.5mil tokens paid for SNOs
so yes it certainly can’t be 4x and i’m sure a lot of other stuff is included in that, such as test data cost and whatever.
maybe we can get StorjLabs to elaborate on what the approximate cost is of the edge services / gateway
because whatever that number is, would indicate if moving most of the gateway traffic to a SNOs cluster gateway solution would be a way to move more earnings towards the SNOs which could host such things.
this is one of the reasons why Storj DCS isn’t directly comparable with stuff like Backblaze, AWS glacial and other similar backup solutions.
and why the past excessive lowering of customer costs for Storj DCS services wasn’t a great idea, something which SNOs will soon be paying for with the new proposed payouts.
2 Likes
Ahh, ok. Thanks for clarifying. Would have been nice to have that data. I was hoping you found something I missed.
I don’t see a way to solve the trust issue here. Customers need to be able to trust any gateway they use and nodes are inherently untrusted entities in the Storj network.
ah right that makes sense… hadn’t thought of that
I think the only reasonable way is to charge for the use of the gateways in addition to the normal prices. The point of Storj was to use the cheaper storage and bandwidth of nodes. With the gateways, Storj might as well just store the data there, because now they pay for ingress and egress of customer data if the customer uses the gateway.
1 Like
Don’t know if it’s worth mentioning it, or perhaps I’m looking at this too simplistic, but if I’ve read correctly about how the satellite infrastructure works, is that it is possible to scale horizontally (cluster of servers) for one particular satellite no?
Well, dedicated servers hosted by Hetzner has unlimited traffic for servers with 1Gbps network card at a relatively cheap cost (the AX161 beast as an example for a price of 145 euros per month). I don’t understand (nor have seen a number on paper) why the S3 gateways would cost so much that it’s becoming prohibitive to scale up the demand.
Edit 1:
Or if we don’t need that much processing power but just a big fat bandwidth, we can go even cheaper but with the same bandwidth per server.
snorkel, Bravo! exactly, the "mantra" of "use what you have" is obsolete, by design!
@snorkel Bravo! exactly
if You pay for renting that “hardware that is running anyway.”
You instantly making incentive to expand hardware, to be paid more.
So ppl buy used HDD for STORJ EXCLUSIVELY to add to theirs setups.
Or even the whole setup to run them.
AND even if not, with money got from You,
they expand existing hardware for them and for the storj both.
So You got to understand that, actions make a reaction,
Any money for the “hardware that is running anyway.”
making ppl want more of that money.
So and to expand and expand, as much as that money allows.
If You think, You gonna just cut the money by half,
so the ppl will be back to using “hardware that is running anyway.”,
Then no, ppl will not use STORJ at all then.
Like many noticed here before,
don’t want to repeat, they explained very well why.
Too little reward, for the users effort, for resources (space),
and for wearing “hardware that is running anyway.”
And one should NOT share a HDD with Storj node, just not optimal.
The HDD should to be only for Storj project.
But do what suits You best.
It should be profitable even for small operators.
By small operator i mean, all You need is to put minimum a HDD, dedicated.
Either by some raspberry pi, or adding a HDD to Your home server setup.
Effort just has to be small ENOUGH, and easy to make a new Node Operator,
so the network gain power.
To people in STORj:
You are not another hosting company, that HAS to be careful if it charges enough for each GB, because it is very limited by the size of the Internet connection in the facility.
so USE it for Your advantage!
Use for Your advantage, what makes You different.
or You won’t break through the hosting market ceiling.
So proposition is:
payout to SNOs:
$3-4/TB for storage filled
(for bandwidth, it depends)
Offer to Customers:
$3 to $4 per TB for storage
$0.2 to $1 per TB bandwidth
That eliminates the high bandwidth cost for Customers,
so eliminates economical barrier to use STORj.
And making it unmatched for any traditional hosting competitors.
Also no obsolete centralized technology, what for, Your not one of them!
$0.2 to $1 per TB bandwidth is small?
How about from 5% of traffic of the entire internet!?
Or from 10-15%+
How fast You can take over the market with this offer?
Details, click to unfold
You can offer flat for everyone, or threshold.
$0/TB if 0 - 150GB/month
(150GB would be $0.03/month, i don’t know, if should be free,
user would need to load some STORj to the acc!
Would be incentive to get STORj from the market, good for price.)
$0.2/TB if 0 - 10,000 TB/month
$0.5/TB if 10,001 - 100,000 TB/month
$1/TB if 100,001 - infinity TB/month
The less TB, the less $, because those are starting business, that needs every penny.
And big traffic, presume, are able to pay finally.
So You can adjust prices later on to piss them off, like everyone does.
I would like it to be flat for everyone.
But later if You need to rise for big ones to $2 or $3,
You will have to for everyone too (small ones).
Or introduce those threshold then, and just rise for selected usage, when You need it.
Also with threshold, You can pay SNOs, for egress too,
from better paying ranges.
If You getting $1/TB, pay SNOs $0.2/TB
Just an example.
Come F*CKING on,
You have a hen laying golden eggs,
and You about to butcher her to save on food she eats.
In this plan:
Customers want to use STORj, because its offer is unmatched.
The more customers use STORJ, the more HDDs are filled,
and SNOs would want HDD’s to be filled fast to get money.
Payout by storage gives stability, predictability of income,
because no one can ensure the egress.
And frankly, there are no storage, and no egress anyway, because no customers,
under current offer $4/TB storage and $7 bandwidth for them.
And bandwidth is the true power of SNOs and STORj, the decentralized internet connections!
So use it as it should be.
You are not another hosting company, that HAS to be careful if it charges enough for each GB, because it is very limited by the size of the Internet connection.
And Energy prices in Europe are doomed to go up and have crisis.
Limits for energy are being put,
like in Poland, basic limit with $0.22/kWh price is now 2000kWh for a year
You pass that, prices goes 2 times more at minimum. (1 USD is now 4,47 PLN)
So You need to pay by storage.
No fake egress needed just to pay SNOs.
You can JUST pay.
Also Fake egress is waste of resources as @vovannovig is hammering.
And makes just confusion for everybody for whats really the networks usage.
Eliminate it!
This plan also:
Entangles SNOs increase with Customers needs for resources.
So the Customers come, SNOs provide Customers, that regulates SNOs quantity,
if HDDs filling fast, SNOs adds HDDs faster (as a new node) or new SNOs come.
If slow, then they don’t add, and some SNOs quit.
The more customers, the more it’s visible to Prospect customers,
and more they want to becomes one.
Also no need for any custody, held for 12 months,
or partial payment, in fear that node will disappear.
if SNO disappear, he will not get it’s monthly payment.
It’s counter intuitive to disappear if You want money!
It’s simple, and it’s easy, SNO gets full pay from the start.
And he won’t get it, if he quits. Simple story.
No playing around with some silly “gracefull exit”.
You can keep it, if You need it.
Just make it as additional reward.
Like SNO has 3TB storage and wants to quit?
Quit with grace, and get e.g. $9 more, or just quit and get $0 more.
Some will just quit, and so what?
Some quit now as well, network didn’t collapsed.
But it MIGHT collapse, if You cut SNOs payouts.
i explained more in length here click me
It also addresses the /24 ip limit problem,
eliminating SNOs die off, at HDD failure,
Because SNOs will have money for HDD replacement, no matter what the egress is.
@littleskunk, click to unfold, about cheating the approach by restricting upload
So in paying for storage model, i described, STORJ should check if SNO is giving enough upload speed for the amount of TB he stores or even want to store. But not just a speed test. Data from real performance. Can satellites know in what time, what sized pieces are transferred? or determinate it somehow, so STORj can know the real speed in action? and if that falls below some point, average, for last e.g. 24h …
if SNO assigns e.g. 18TB to the network, should have some reasonable big enough up speed, he shall be informed before, what are the requirements.
if real performance shows SNO doesn’t have it, SNO would be informed, and some action need to be design for that.
For example if You assigns with 18TB and really your pieces goes with 10Mbps upload max, that’s too slow, either You rise that somehow, or get penalized, paid less per TB storage, or something else.
Just some Anti-cheating mechanisms so ppl won’t just cheat.
Notice, You could eliminate VPNs imitating another location with this, because cheap ones, rarely got more than 20-30Mbps upload, and making STORj bad latency. No more nodes siting in data centers, imitating real user, and stopping decentralization.
SNOs still should have options in config file to limit monthly TB bandwidth,
and
to limit requests per second.
So the connection won’t saturate, making internet hard to use.
You see this is whole new approach, but i believe it’s critical to take it, to prosper,
And the details how to execute it, should be figured out.
You didn’t responded to my previous message to You,
i think this here is shorter to read now,
and You can see, it address all 3 points You mentioned in Your previous post.
- Customers will pay more than SNOs gets for it. Solved.
- Can be free accounts, just if they don’t pay up to 150GB, SNOs don’t get paid for that too.
SNOs have a lot of bandwidth to give away. But You have to somehow make sure it is not abused, like someone said: let add customers phone, or credit card. But that i don’t like, because i like no knowledge, no KYC, especially not right from the start. When i need to just upload something for someone fast, STORj comes handy. And if someone uploaded up to 150GB backup once, up to now, it should be there free for ever. "Lex retro non agit " Or You just limit it in time for testing, from now on, and after e.g. 3 months time out, pay us. Or, there is no free accounts from now on, and traffic is $0.2/TB from first byte. That would be $0.45 for 150GB stored, for ever and Armageddon resistant, and $0.03 for every 150GB bandwidth. Solved. - Test data elimination. No need for such test data as it is now, just to pay a SNOs. Solved
7 Likes
I strongly suggest that Storj does their very own homework before - to quote @John -
grinding their SNOs to a pulp.
Suggestions on cost cutting, revenue increase and even marketing and targeting customers have been made. Many of them.
I fear that the proposed lower ends of the range ($0,75/$1,50) (storage/egress) will not work for me. Operating nodes at a loss without prospect of positive returns would mean to shut them down immediately even without graceful exit. To me it looks like that with these low proposal node operation is only viable at scale, which will lead to fewer larger nodes and more centralization. Because with smaller nodes your earnings will not be going anywhere.
For now I cannot tell what my actions would be regarding the proposed ranges but quitting node operations completely certainly would have to be seriously considered. It is becoming clear that the Storj mantras to not invest and to use only what you have are going to be no longer true for a viable operation.
Payout proposal
| Storage | $1,5/TB | I strongly speak against lowering current storage payout price because if a node is full egress drops significantly and the SNO is basically left with the storage payout. This makes it less viable to run smaller nodes. Also storage costs seem not to be the major issue. So $1.50 is and should remain the absolute minimum. It should rather be raised than lowered which could open additional room for lowering payout for egress. |
| Egress | $10/TB | For regular egress and repair traffic. Personally, I would be even willing to waive the repair fees for Storj, however it is true that if there is no price tag for it, there is no incentive to use it economically. |
Alternate option
| Storage | Egress | Repair | |
| $2 | $2 | $2 | My calculation claims that this could work, yet this would have to be verified more thoroughly. This payout scheme would reflect one idea that I had in the past, as to why distinguish between egress and storage at all but also puts the focus on reducing egress costs as this is where the actual economic model fails the most and causes biggest losses. |