1 PB node(s) as fixed money income

I have some small and medium-sized nodes ( around 50 TB ) and thinking of creating some big nodes as an investment.

The main idea is to use 3 nodes with 24 hdd filled with 20TB drives, each using 4 x Z1 6-wide raid scheming, getting around 360 TB of usable space each or 1.08 PB in total.

This whole setup will cost me around $25k, and I expect the payment mainly comes from the space used ( $1.5/TB * 1000TB ). Everything else will be a bonus, or used to pay the maintenance.

The energy usage ( around 800W ) will not be a problem, and I expect no more than $50 per month. And I have 8 Gbps fibre Internet ( yes, it is 8000 Mbps ).

What are your thoughts on this? Is it feasible or not?

There are certainly 1PB+ setups out there: so what you say is possible. But when you say you have some 50TB’ish ‘nodes’ already… do you mean working Storj nodes… or do you really mean ‘servers’?

Because buying the gear and getting it online is straightforward… but you only get paid for the space you fill… and that can take a long time. And the only way to really speed it up is to use more different-/24-subnet IPs. If you only have one IP… you’ll never fill all that space.

(Edit: Do you have one of the now-unavailable Bell 8G fiber connections? Jealous!)

1 Like

$50 per month is certainly doable (thought you mean $500), the problem come from /24 rule of storj, so you would need more ip address than storage. Let do some calculation:

50$ mean about 33.33TB, so without investing anything, your current 50TB node should be able to give you that (but because of /24 rule you need to spread it out on multiple nodes).

Your 8 Gbps won’t mean much (unless you are on Understanding the Commercial Storage Node Operator Program - Storj Docs).

Hi Roxor,

yes, they are nodes with sizes between 30 tb and 8 tb.

and again, yes, I have bell fiber at 8bgps… i was a bit sceptical when they offered it, but now i can even go up to 10gbps on it.

I started a new 30tb node in the last days of May and it filled up in a month, using around 200 mbps of bandwidth near constantly around the last month.

I have only 1 IP, but I think I can get another fiber from Rogers ( 1.5 gbps ) and have the 2 ISP connected to my firewall to load balance the nodes.

Remember that the firehose of test data we’ve seen lately has been extraordinary… and perhaps short-lived. It was taking people 2-4 years to fill 8-12TB HDDs previously.

(But if you already have most of the gear: there’s not much to lose giving it a try. I just wouldn’t drop $25k in fresh cash on Storj. I’d start with what I owned… and add drives as they fill)


Hi kocoten1992,

Sorry, for $50, I meant the power/month cost ( based on 800w/h ) of this setup.

I expect around $1500 to $1800 per month in income from the storj nodes.

Thank you for point me the document about the Commercial node operator.

Yep, and that’s likely about all you’re getting too. The recent data, while trying to be representative for prospective customer behavior, has been uploaded with a 30 day TTL. Meaning it’s being deleted as fast as it’s coming in now. You’ll still get data from other customers, but it will be much much slower. DO NOT BUY $25K of hardware on a single IP. You are never going to hit your income targets. Not even close. Maybe if you cut a 0 of those numbers and have the patience to way several years to hit that target. I’ve been around since launch of the network and have used 1 IP for the longest time. Now recently I’ve been using between 1 and 6. And I make less than $100 a month atm. (Though it used to be around $150 and it can certainly get back up in that range again.)


I would start with 1 node (360TB) and when it is filled… add another one. At the moment there are many uncertainties and test data that can be turned off from one day to another. Wait for signed contract announcements.


I have 2 nodes on 2x22TB drives on the same IP, and I can confirm that the TTL data ingress has reached a stable state. I took in 32TB of test TTL data and it’s almost a week with stable state. 16.5TB of data on each node. The nodes without test data have seen a growth under 1 TB per IP last month.


Off-topic: but you obviously have the skills… and it sounds like you’ve had the opportunity… to become one of the largest SNOs. Have you ever had the urge to expand a bit more aggressively: in the times you did see above-average uploads?

You could have been Th3BrightSilence by now :wink:

[Off-topic: Storjlings can skip this post]
Not really, no. My current hardware is reaching it’s limits, so I’m fine with expanding slowly. I’m probably going to get rid of some IP’s because the new node selection makes them give diminishing returns after like 2 or 3. And I never bought a server/NAS for Storj alone. Only HDD’s and an external USB connected 10-bay device. But let’s keep this topic clean if Storjlings are going to look at this on Monday. I don’t want them to have to wade through off-topic messages when working on solving the issue.


Yep, I pretty much figured it out. I was hoping to use my fibre bandwidth for something useful. I will start with the first 6-disk raid z1 and add more as soon as they get filled.

Thank you for the considerations

yep, as I said, I will start with 6-wide raid z1 ( around 90TB free ) and add more 6-wide raids as they get filled.

I hope some big shots come to storj and I can fill they up really fast.

What can I do to discover if the used space is part ot the test data or something else? Sorry, I am really new in the Storj.

I wont suggest you to run so big nodes like 90 tb, garbage collector, used space filewalker will last forever on this one.

Test data comes from the Saltlake/SLC satellite: which uses the pmw6tvzmf* folder under files/config/storage/blobs. So the size of that folder (compared to the other 3 for US/EU/AP) will tell you how much test data you have.

I bet if you check your 50TB of used space on existing nodes: 99% of it is from testing.

You can run some command to scan the size of the Saltlake folder. Saltlake has only test data, and the most of it is the new TTL.

You realize that you have no contract with Storj?

What do you do, if they give you only 10TB total stored data next month?
Does your setup still work if you only get 15$ a month?

  1. This is not investing, but spending money on toys
  2. You will not get the expected result
  3. If you are ready to spend money without reading at least this forum over the last year, then this is free money and you are ready to risk it - then No. 1 is “buying toys” …

In short, your idea has no prospects.


Go slow and steady. Wait until your current nodes are almost full. Then buy your preferred hard drive bay and add one drive at a time. Wait till that’s full, and repeat.
I will be doing this once I’m totally full.