I was thinking of a hopefully feasible offer that Storj could build to attract node operators. First, assumptions/constraints:
The offer needs to be at least minimally profitable for Storj. This means covering the human cost and any satellite-related costs. For example, we know that the number of segments stored in the satellite’s database has an impact on costs. API calls aren’t free either.
We can use RS numbers of 29/46/54/70, meaning the expansion factor of ~1.9.
The offer can be tailored to incentivize node performance.
The offer needs to be truly limited to node operators to specifically incentivize many people joining.
So, what I’m thinking of right now is a following storage offer:
Storage cost: 0.0002 USD / segment / month, 30 days minimum. Just that, no additional fee for size. At a maximum segment size of 64 MB, this comes out as 3.125 USD / TB; and directly incentivizes having a small number of big segments.
Egress cost: 0.0004 USD/segment, that with 64 MB segments would come out as 4.7 USD / TB.
And an additional cost for metadata API calls: I’d start thinking of prices comparable to AWS, e.g., 0.005 USD / 1k requests.
API limits depending on node speed. The faster the node is (in terms of races won), the more requests per second you get. Not sure what would be a good baseline number here, but I’d try to set up the floor so that slow nodes can download at roughly 5 MB/s, allowing hobby video streaming.
No S3 gateway available. If you want S3, you run rclone serve s3, or the Storj gateway yourself.
No support except for a subforum here.
The offer needs to be at least partially paid with node revenue by using Storj token payment address as the wallet address in the node itself. And the offer only covers the total bytes-seconds stored across all nodes using that payment address times the availability factor computed as: max(30%, 20 × ("online score weighted with node/satellite data" - 95%)) (or a similar formula—the key being, there’s a floor to have a safe value for everyone, and only truly available nodes get anything above the inverse of the expansion factor). Anything above that is automatically charged as the regular Archive offer. And on the other side, the spare bytes×seconds number can accummulate over several months to allow some peak usage or one-time longer downtime.
As for the segment-based fees—you want “cheap”? you use rclone’s archive backend to combine small files together. Or use restic as an intermediate layer. Or whatever else client-side solution you find. The key thing is, Storj does not need cover the extensive costs of metadata, you use a client-side solution to make it cheap for Storj.
With this structure I would expect that the network would attract many hobbyist data hoarder/home lab users, gain a lot of mostly static data, large segments, that on average are rarely downloaded, hence driving the average segment size up, and reducing the average deletion rate and egress.
This works out to around “get 1 TB of storage for each 2.1 TB of stored data”, covering the expansion factor with a bit of a margin for Storj to cover costs I don’t have data to estimate. And whoever does not actually provide either the amount of storage to even cover the expansion ratio of their own data, or high availability that could help reducing the RS numbers, does not get to leech more that their fair share.
Question to node operators: would this offer work for you? I would myself move away from my current backup storage provider with my 15 TB of archives if this was available.
For me, the math is pretty easy. My equipment is powered on anyways, so weather or not I’ll store additional data boils down to one thing: Is it profitable.
If running an additional (lower?) tier of StoJ would still provide me with a net positive return, then I’m all for it. Honestly, I don’t really care that much for all the details - it’s the number at the end of the month that’s the interesting part to me.
Storj is focused on business/enterprise customers: who compared to regular individual users… spend a lot and need minimal support. I think a flood of new users that use less, pay less… and need more support would kill them
Indeed. Right now per Storj Pricing | Seamlessly scale on the cloud built for media production Storj has three tiers: Regional, Archive, and Global. I’m suggesting a fourth one, limited in terms of audience (only node operators) and scale (only the rough equivalent of what they contribute to the network), and leveraging the same public network of nodes—and the same payments, same storage/egress payments to them—as the other three.
The data I have suggest that this should be feasible. I posted it here to estimate interest among existing node operators, because that would be a proxy to estimating whether that kind of offer would attract new operators—the primary goal here.
So please vote if you would be actively using this offer, e.g., for your backups or other storage needs—the limitations of the offer, and pricing scheme are acceptable, would be attractive to you specifically. Not just if you see the offer valuable in general.
I guess I’m with snorkel on this one: Storj has zero need for additional SNOs. They shouldn’t be offering any less-profitable storage tiers to attract providers that do nothing for them.
(And SNOs can already provide enough local space… to cover their Storj S3 bills?)
TL;DR; I vote no: a new tier like that harms Storj more than it benefits me: and I want them to remain in business so they can keep sending payouts
having to install software instead of using integrated S3
having to pay for egress and API calls
just so I pay 3$ instead of 6$ per TB.
Or to frame it differently:
If I pay 3$ per TB, can pay with CC, pay 3$ per TB download, nothing for API calls, and get a TrueNAS integration, I would consider it.
Why? Node operators are paid in money. What the point in also paying them in [subsidy for] the product? Or entangling consumption of storage with providing of storage in the first place?
Why does storj needs more, let alone many, people joining? There is already surplus. The control knob is a payout and/or artificial traffic, not discount on the product.
I’d imagine people who provide storage and need cloud storage are very weakly overlapping sets, if at all.
Not related to any topic in particular, but individuals, I imagine, don’t ever use a full TB for cloud backups of important personal stuff. You could fill a TB only with photos and videos, but those are usualy backed up on sticks and external drives. I believe one would only pay for cloud for critical stuff, docs, projects, dbs.
I have the opposite experience. All of my data is in iCloud, including photos, but there is about 1TB of stuff that are not photos or videos.
That’s a working set, with on-demand sync to places where data is used. It’s replicated to two storage servers, with versioned snapshots as a backup.
I would not consider flash drives or solitary hard drives safe enough to even bother writing anything to them, other than for faster transfer — but that almost never required, Ethernet is fast enough. People who “backup” to them underestimate the risks.
Separately, I consider personal stuff more important than any other data. Personal stuff is irreplaceable.
Hello, where are Storj’s business customers? I don’t see any growth, not even in the last 2 years.
I am always open to something new.
Due to AI, data centers, market prices for hard drives are skyrocketing, electricity in the USA is getting more expensive. I always thought that slowly the global network must be growing. But it isn’t. Maybe the project has already failed, I don’t know. …
I hope not. Lately, you haven’t been getting any information here either.
It looks like they added about 7PB of customer data so far this year. But it also looks like the strongest growth was on the US Satellite: so perhaps nodes in Germany didn’t win many races?
We have almost no voices of people using it this way so far on the forum. My hypothesis is that Storj offers are not useful for them to justify the higher price of the service. This is one reason I made this post—to check whether there would be interest in using a trimmed down and cheaper service.
From Storj’s perspective, that would be one more control knob. Maybe one that would them, let say, reduce RS numbers even more.
Storj is a perfect piece of technology to implement 3-2-1 backups. You store your backups locally, and just by getting bigger drives, you also finance remote storage. We’re not seeing this though, see my hypothesis above.
First of all at the current utilization rate we certainly do not need even more SNOs at this point.
And for such an idea to work, I believe the process must be more user friendly.
I don’t want to deal with worries that my earnings might not be sufficient and have my data deleted for whatever reason nor constantly monitoring or changing payout addresses and else.
Normally there is bank accounts and credit cards for that where you can receive payments and send payments.
Storj would have to change the process in a way so that when an amount is due, they check the erned tokens first and deduct it from that. And depending on the rest, either charge the credit card or send the overpay to SNO or keep the overpay as per SNOs liking. Maybe even a way for SNO to set a limit, like not more than $10 at Storj, rest to SNOs wallet.
As SNO I would not want to deal with the hassle to constantly check balances, change payout addresses and worry if my earnings taking the correct route so they or my data don’t get lost.
And also: Storj has lost trust and reputation for their past actions regarding price changes, free tier adjustments and the back and forth with the introduction of the minimum fee.
At the moment I don’t think that we need an additional incentive to attract more operators at this usage.
This already is working like this. The invoice is trying to use credits in this order:
How this affecting trust?
The business should earn money, isn’t it? The promotional credits are over, everything become only higher in costs, so we forced to increase prices too, and I don’t think that pricing requires any trust, but research.
