What is the price of repair traffic?

Node Operators
1

No abstract percents here just works, it’s just ridiculous
I have already said that I don’t have such calculations. They are VERY sophisticated. When I see the exhaust of my scripts from many farmers, I can express my opinion on the basis of these reports and my experience. But it is so trivial that it used to go so far as to prove the obvious was not required.

Yes, I just end the discussion

And taking this opportunity, because I am not going to visit this forum for a long time, I want to raise a question that has not been fully answered.

What is the price of repair traffic?

As a tactful person, I won’t post a discussion here from a chat with a certain Victor from the Storjlabs, where he explained that repair traffic costs 10$ and (by definition) it can only be ingress, and there is no other (egress) at all. But you can make your own opinion about the competence of the Storjlabs leads and their advice.

By all advices from v2 from @Alexey, Next by Viktor, repair was cost 10$ for Ingress. And egress repair will cost as usual 20$. But now we billed 10$ for egress repair and 0 for Ingress repair.

2

May I point you to this button ==> image
It’s on the top right.

Price of repair to whom? As you know SNOs get paid $10 per TB. So that would be being paid for repair. The customer doesn’t pay for repair at all. So is your question the cost of repair for Storj?

I that case, to repair 1TB they would have to download a little more from SNOs that still have pieces of the file. 35/29 to be exact. So 35/29 * $10. Plus the cost of compute on their satellites which are hosted on cloud platforms as well for recreating 55 missing pieces. I don’t pretend to know what their costs are for that, but the compute part is probably not that high. Then egress costs of that cloud platform for 55 pieces. Again, don’t know the exact cost, but they use GCP, so what the hell, lets look it up. Looks like they charge 0.08 USD per GB if you use more than 10TB per month, which I’m going to assume Storj does. So, we can add another 55/29 * 80 USD to the repair bill for that 1TB.

So… 35/29 * 10 = 12.07 USD
55/29 * 80 = 151.72 USD
163.79 USD + compute cost per TB total.

With the repair threshold at 35, 55 nodes need to disappear before a repair gets triggered. However, offline and suspended nodes count against that number as well and I believe the current repair threshold is set quite a bit higher. I think around 55, but I’m not sure.
So somewhere between $3 and $6 per TB per node lost?

My held amount more than covers that, but it should be said that the only reason it does is that most of that was paid out at significant surge percentages up to 500%. If there wouldn’t be any surge payout, my held amount would have been about 45 USD (this can be calculated using the nodes db, so I did). Since my node holds 10.7TB currently, that seems right in the sweet spot to cover the costs. However, half will be returned to me soon. So after that my held amount won’t cover the full cost of repair had they not been so generous with surge payouts. I imagine this is the calculation Storj did as well to determine held amount and payout to SNOs for repair.

Does that answer your question? Or is it too theoretical again?

I didn’t even have to coyly refer to private conversations with insiders. All of this is public information. A little elbow grease and the information is right there.

V2 really didn’t have any repair to begin with, so I’m not sure why that is even related. And replication was done by the nodes, not the bridge. So the “cost” would be egress on nodes and wasn’t paid at all at the time. Forget about V2, none of it applies anymore.

3

V3 discutions starts in rocket chat in V2 topics before this forum started. Where was first mentioned price for repair as 10$ Ingress & 20$ egress.
In you calculations you forget the fact that escrow is not required to accurately cover the costs of a particular farmer. This is where the analogy with insurance would be acceptable, and not where you shove it. In other words such calculations not acceptable for escrow. Where must be calculations of probabilities.

It is unfortunate that you wrote so much text in vain, especially in response to a rhetorical question.

4

I thought you promised to leave. Please do.

I put effort into calculating an answer to your question and this is what I get?

2 Likes
5

Thank you for your attempt. Just multiplying and dividing the numbers is not enough to count something. You need to understand how to count.
Insurance is considered different. I am sure you will be able to understand this by trying to reconcile your calculations with a repair, say for 16-28 months. You will not divide by zero? After all, it is precisely 0 that will come to escrow from 15 months. And it’s not zero that those crumbs that remain in escro after 15 months will never cover repair traffic, say if the node grows significantly in size

6

Could you point out a quote where it says?

7

If somethere worked rocket chat with v2 conversations?

8

I think it was shutdown a long time ago.