So as I was looking at my current stats for both my nodes i was thinking about how the current low ingres must also be impacting storj with reducing the amount they have in held back to cover their costs for node churn.
My oldest node is in month 7 and hasn’t even reached $6.00 in held back yet but it is well over 1TB stored now. Even after the recent deletes… Given that some of that is also from test data it means actual customer income to pay for node churn has to be quite low still and if there will be a significant number of SNO’s leave that increase that has to impact storj’s reserves.
I know @BrightSilence has said SNO’s leaving isn’t significant for the stability of the network and that may well be true - but there are still significant costs associated with SNO’s leaving. I also wonder if storj has any statistics on average age of nodes leaving the network?
It seems this is a bit of out of context paraphrasing. I definitely wouldn’t put it like that. Node churn is something that is a concern. Storj wants to do as much as possible to limit that, simply because it’s costly. I’ve said in the past that the current levels of node churn aren’t concerning. And that node churn can never lead to endangering data reliability. There are too many protections in place for that to become an issue. But you still want to keep it low to prevent high repair costs.
I would say the current held amount structure isn’t ideal. Nodes grow slowly at first and don’t make much money. The way held amount works now, most of that small amount of money is held back. That still doesn’t really help Storj and barely provides an incentive to stick around. While later on, nodes accumulate a much higher potential repair cost, but no held amount is being held back anymore. I feel like held amount should be something that is kept at a certain level. Say about 2 months payout. With a maximum of 50% of a payout being held.
So say month one you earn $1, the goal would be to have $2 held based on those earnings, but we never hold more than 50%, so you hold $0.50. At some point your held amount will be close to 2 months earnings. Say $20 when you make $11 a month. In that case only $2 is held and you would be exactly at the desired level of $22.
After 15 months you could drop that desired held amount level to 1 month worth of payout perhaps if you still want to give loyal SNOs some compensation at that time.
3 Likes
The repair costs are based on a certain amount of data to be repaired so why wouldn’t the held back amount be based on that too - at least in the longer term?
I would agree with this.
Agree too. The current held amount is barely a deterrent. Perhaps in the early days with higher ingest and surges, it was.
Could you ask SNOs for a deposit and the arbitrarily ban them? I don’t think so.
How about new node’s get surplus copies of data only, so no repair is needed: but then the uploader pays for the noobs repairs
Perhaps unvetted nodes could only get block 81?
What about if instead of turning off the held amount after 10 months it was kept on at least for 2-3% until the reserve built up to 120% the cost of rebuilding that node? That extra could then be applied to nodes that exit far earlier in the process? Now, that equates to older nodes subsidising the new players but given current ingress levels I don’t see how you get any more money out of newer nodes.
I thought about something like that in the past: Minimum Threshold for Storage Node Operator Payouts - #146 by jammerdan
But if repair cost are really a problem, than Storj could always change what they pay for repair egress. There is no reason why they have to pay $10 for repair downloads per TB. It could be more or less (even nothing).
1 Like