I am a cloud noob, so I might be wrong… But I would not expect big players to have that much unused space. Wouldn’t they have a contract with a HDD maker or distributor to get whatever the want to add within 24 hours?
it is what I found from google search, let’s see if anyone who participates in datacenter capacity planning chimes in.
But either way,
- there is always a gap between when storage is added and utilized (you can’t utilize storage until it’s added), and hence, there is always unused storage available.
- running close to 100% capacity is risky, and expensive, because you won’t be able to absorb fluctuations, and will have to drop customers, and since expanding capacity is not an instant process, you want to give yourself some buffer. That’s the buffer storj can use.
- The argument above still applies, if you grow by 2% annually, you would not want to go through trouble of replacing disks annually, so you’d add 10% now. Those extra 8% storj can use.
Using (temporary) overcapacity doesn’t fit the 1 node per disk recommendation nor the hold back requirement. This was always fairy tales, storj is not build like that.
Storj accepts all variants of a not used space, from datacenters, homies, everyone who want to participate. We almost constantly growing.
However, it is likely not designed for the setups built only for Storj, because the usage is not predictable, it always depends on the customers no matter what.
It doesn’t fit because you can not easily reclaim space without losing payments. Not a problem from storj’s point of view.
Well, that’s how renting business work. You landlord stop renting out your house and move there yourself, you stop getting the rent payments. It does not mean you should not rent out your house if you don’t need it for years
In my country the landlords asking for a held back not the tenants.
If you’ve seen that bulk data would be deleted, and your geo has recently been a focus… I’m not sure what your question is. You know nodes are having data removed… and have come to ask if anyone else has noticed data removed?
Yes. Yes we have
Don’t sweat it: we’re all living through the test-data-TTL-churn right now, so ingress has been lumpy… but underneath it all is the same natural customer growth slowly pushing used-space numbers up. Your nodes will recover.