I am a home miner with a small rig who got interested in blockchains technologies 2 years ago, and I kinda like the idea of decentralization so I became interested in Storj and what it has to offer to the world. I am planning to make 1 single investment to purchase a 24T HDD to run it along my gaming/mining rig, but first I need an advice on what HDD to buy. I want to know about what specs to focus on, best brand (durable product), best price, etc.
By the way, I have free electricity so any HDD which consumes more kwh for its price related must not considered as turndown.
24TB will take a very tong time to fill. You can get a smaller drive, 8TB or so and it will still take a long time to fill.
If you want to run a single large node, consider using RAID - you would not be happy if your drive fails and takes the data that you had accumulated over 3 years.
Instead of RAID, you can run multiple smaller nodes on separate drives, this way if a drive dies it won’t hurt as much. Also, this way, you can buy one drive, fill it up, then buy another.
Do not use SMR drives.
What size drive to buy initially?
I have run my node since before the last network wipe and other than one week it always had free space. So far (probably over 3 years) I have about 21.56TB of data. This is pretty much the maximum possible for an old node. It will take a very long time for you to get that much data.
My node currently is filling up at about 700MB per hour, so 1TB would take about 2 months to fill. A new node (after the vetting period) will initially fill a bit faster, but not by much.
So, you could probably buy a 6-8TB drive and it will last a year, then you can buy another drive and start a new node.
The recommendations here are valid. I just wanted to point out that node operators “may” see faster adoption of their unused storage as Storj continues to gain new customers. The more we add, the faster usage grows and data becomes more fluid.
In other words, drives may fill up faster than they have in the past. Lots of variables there, but I think ingress will soon see an uptick that is noticable.
That’s always great to hear! I’ll keep an eye out for that for updates to the earnings estimator.
That said, more data makes it more attractive for node operators which leads to more nodes being launched and the ingress being spread over more nodes. There is a kind of self balancing in play here. Also, when more customer data gets on boarded I think we may also start seeing some deletion of test data. As you said, many variables.
Either way, I think we have good times ahead. Keep the customers coming!
@tolonuia: The earnings estimator I referenced can be found here: Realistic earnings estimator
It gives a good estimation of expected growth for nodes based on recent network behavior. But of course if ingress increases things may start to look better than it currently suggests. I tend to advise people not to assume additional growth, but see it as a nice bonus when it does. Either way, if network behavior shows consistently higher ingress, I will be sure to update the numbers to reflect that.
Im curious about the effect that removing test data will have on $/TB numbers. Do you think that will increase it? Are we paid anything for test data and does it get read ever? Or just fill space?
Yes to both. I believe they stated the intent to simulate customer behavior with it. So I don’t expect it to shift the numbers much. But if it is removed in favor of customer data, it’s a fairly significant change, so I guess we should expect some impact. Hard to predict which direction that goes in though.
Saltlake is test data. Europe-north-1 is test data + some Storj internal projects. US2 is used as a test satellite to test new satellite features and has been used for public testing of those new features as well.
Saltlake and Europe-north-1 are quite big, but haven’t seen significant ingress in a long time. All ingress we see now is from customer satellites. The balance between test and customer data is shifting fast towards customer data. I don’t know the exact totals at this point, but the trend is very promising.
Edit: Actually the link @jammerdan shared has the stored totals per satellite
Yeah that tracks. Those two sats are 51.3% of my stored data. Glad we’re moving toward more customer data. At the beginning of the year it was closer to 2:1 and now its 3:2. The right direction.
I think I am going to leave what I said as is. After almost a decade in this business, I know enough not to say more than is needed. If the business wants to make any announcements, they will.
As for test data and how long it will remain on node’s drives. That may be better answered by @littleskunk and @stefanbenten
Well it was mentioned before that it was partially there to be able to quickly get verified free space on nodes in case of a sudden rise in demand. So I was just going off that. I’d say it’s a fair guess to assume it will stick around if available space rises with demand and probably be partially removed when demand suddenly outpaces supply.
Since the beginning of the year I also have seen an increase in daily data filling. It varies from 20GBs/day to 30GBs/day in 10GBs/day increments. Even saw 40GBs one day.
Based on several screenshots of my node since the beginning the year 2022: There isn’t a major increase in the Ingress+Egress bandwidth for any of the first 6 months of 2022. (This particular node is located in Europe.)
But there is a large difference in Ingress bandwidth depending on where a node is located (such as: Europe vs USA).
