Is it possible to host multiple nodes per IP address? For example, could I grab a 4-bay NAS with a quad core CPU and create 4 virtual PCs each with its own RAM, CPU core and 20TB hard drive with a sufficiently fast network?
Also, if I have a node that’s been running for two years and I add more nodes to the same IP, are those new nodes trusted to store data at the same level of trust as the first node? i.e. is the amount of data pushed to the new nodes equal to the amount pushed to the old one?
Each node has it’s own Identity files… so each will be audited separately… and be subject to it’s own 9-month withholding period. Yes they can all be on the same IP (but different ports)… BUT they’ll all share the same ingress. So unless the nodes have different /24 IPs… you won’t actually fill their disks any faster.
It will work though. Many SNOs let a node fill one disk… then start up a second node on the same IP to begin filling a new disk…
4 nodes per IP is okay, just use different listenting ports. But 4 nodes per disk leads to bad performance , and also violate the TOS.
I was hoping to use a different disk per node for more revenue. It seems to me that the 9-month trust window should be per provider rather than per node. If Storj has decided an SNO is worthy of full payment, why wouldn’t additional nodes by the same provider be trusted equally?
The thing that makes each node unique it it’s Identity files. Storj had no idea about the hardware or config each node is running on (regardless of a common payout or email address) so each is monitored for uptime and durability.
Like you could run one node on the worlds best hardware, and have 100% uptime, and a perfect reputation. Then run your second node on a Casio Calculator Watch that loses every upload/download race 
Because they may be located in different places, connected with different internet providers, powered by different power companies, number of cats that may yank out power cable can also vary, as is the speed of response and fixing issues.
It’s the specific node that needs to prove that it can stay online and not lose data, not the owner.
It is still useful to have multiple nodes in one place. They are treated as one node for ingress, as it should (it does not really matter in how many instance you split the space, it’s the same storage at the same place), they fill equally, so when you need your space back you can just delete enough nodes to get back enough space. Remember, we are renting out the space we not yet need, but might in the future.
Also want to add that if there are multiple nodes per IP, you effectively splitting the load between them. Instead of one disk getting hammered, that spreads out to multiple nodes. The network treats them as one for ingress. For egress, they can also help since you are effectively reading from 4 disks instead of one.
Just create 1 VM and install docker in it. Uses much less resources. Or even better, install docker on your NAS if possible.