So, as I expected I have seen the zksync option being pushed hard and I just don’t understand why.
To give some background , I am not a crypto speculator. I’ve never had any particular interest in bitcoin or etherium or any of the others. I have been interested in distributed internet projects for a long time and have done both seti@home and distributed.net back when they supported a lot of different cpu architectures so I could run something on my Sparc and Alpha boxes. I got interested in storj because of the project itself and was also interested in the ability to earn something from it.
From my perspective the intent was always to transfer to fiat. So the fees to be able to do that are key to me. I originally joined the project back in November 2020 and we moved cities late December and I took the opportunity to re-do my nodes and make some changes. So, I never got a chance to earn anything significant with them (This was why I called my storj earnings worthless in other posts - because literally after any fees the net value was basically zero or even negative. It was not disparaging to the project itself. )
So, I redid my first node in January and brought up my second about a month and a half ago.
There are two parts then that are of interest: 1. How much space could I realistically fill in a year and 2. How much could I earn. Thank you @BrightSilence for your community calculator who basically provides those numbers. I am currently earning about $2.50 per TB. (before held back). I’ve seen here it mentioned this can be as high as 4,00 per TB. So, my question is, how much is the average SNO likely to earn in the first year and assuming he/she takes the zksync option what can actually be done with those storj on L2. I see only 2 options. 1. Coin speculation or 2. Buy storj services. The fees to get to L1 will eat out too high a percentage of the net amount to be in any way meaningful. This of course does NOT apply to the longer term players who experienced surge payments or had the chance to grow their data levels. I’m specifically addressing this to people who have joined storj in the last year.
I currently have 2 nodes. One that has been online since the start of the v3 network, and one since Sept 2020 (8 months old). They are both behind the same IP, but the older node has almost always been full, and both nodes are full now, so I don’t think one is significantly affecting the other in terms of payout.
For my 8 month old node, which is sharing 1.74 TB, I have earned on average $1.61/TB (before held back). For comparison my original node in the same time period, which is sharing 3.55 TB, has earned on average $4.59/TB (before held back). This indicates to me that sharing anything less that 2TB of shared space is probably not worth it, since egress will scale with the amount of data you are able to hold.
Not directly. But if there was a period of time during which data that is now frequently downloaded was uploaded, well nodes that were receiving data at that time will see a benefit. There is one thing that goes up a little over time, which is income from repair egress. The older the pieces on your node are, the higher the chances the segment they belong to might need repair. On my biggest node that accounts for $3.55 out of $25.06 total earnings so far this month. It’s a nice bonus, but won’t change your income massively. Most SNOs make about $3-4 per TB atm.
Keep in mind that they did that to attract different use cases that use more egress. So hopefully that works out and the different type of data we get compensates for the lowered payout. And at least for this year, they promised our payouts wouldn’t change while customers are already paying less.
I’m also aware of the time delay to deployment. However, if you are on the low side of deployed storage and have say just the 1TB used i see the potential downside as being greater than the upside. Still, we won’t know anything for sure until we see the proposed new regime and run the numbers.
Not that I know of. The white paper does include some stats and analysis of node churn based on their v2 experience. But if you’re looking for expected lifetime for your own node, I doubt they would have usable stats for that. Node churn can just be someone leaving and there is really no way to differentiate that from Storj’s point of view.
And while I typed that I realize you’re probably asking about whether they modeled expecting income over the lifetime of a node. I think some of that still applies. From what I’ve seen, they take node profitability seriously. The expected lifetime of HDDs is pretty high these days though and you can easily expect 5 years out of an HDD and often many more. I think doing that calculation is kind of on us SNOs though. Their advise has always been to just use what you have. And I get that. If I were in the driving seat I would probably still aim for about a 2 year ROI. That kind of ensures profitability and gives them something to aim at. But I don’t know if they have a goal like that internally.
Thank you. Yes, my perspective is as an SNO.
As you well know storj does not recommend redundancy as the network can cope with the failure.
However, the SNO would still lose that income from the node failure.
Also, if you do follow storj recommendations and use what you have then the drives already have time on them.
(My second node is based on a 3TB WB Black I no longer use for main data storage as it has over 60 000 power on hours now). But, I’m “using what I have”.
Of course some drives just keep on chugging. Heck, there are still MFM and ESDI drives out there that still work.
I still have 2x 320GB HDDs that are still going. Yes, some HDDs seem to just live forever. I actually still have a 160GB that also works, but not currently in use.
Yes, but the money is already spent. Anything you make of a drive that would otherwise do nothing or of free space that you’re not using is just bonus.
@penfold Brandon from the product team here! We do some modeling on the average expected lifespan of a node in terms of node churn. We need to understand our node churn and how long nodes stick around on the network to determine if the reed solomon numbers we use will achieve the level of durability we have in our SLA. It also helps us understand the natural decay for a file and when new files will start to need repair. repair is a very costly thing to do which is why we incentivize long-term nodes!
part of the reason we don’t recommend redundancy on the node HDD level is that we believe it’s more profitable to run 1 node per HDD. If you have 3 HDD you want to dedicate to the network and you run 1 node per HDD without any redundancy then if one of them goes bad and that node gets DQed you will only lose 1/3rd of your total held amount but the upside is that you are storing a lot more pieces of data for a lot more files so it’s more likely your nodes will have more bandwidth used which would yield to a much higher payout!
Hi @brandon ! Thank you for taking the time to reply. What is not taken into account in what is written above is the time to replace the data based on the current and recent low levels of network usage. If an SNO has for example, just a 6TB node and it fails due to drive death then the SNO is going to need a very, very long time indeed to replace that data. Perhaps longer than a year. If the network was heavily utilised then yes, for sure the storj position makes sense. But that isn’t the position we are in at the moment. Oh and the total held amount isn’t even factoring into my calculations. I’m just ignoring that.