I can’t stop thinking that Storj can have a better environmental footprint than centralised cloud storage. Yet I don’t see Storj making any communication around this so maybe I got a bit excited…
My initial hope when I started to follow this project back in 2014 was that this could really act as something similar to Airbnb or Blablacar for cloud storage, with the same underlying objective to decouple growth from the use of resources. However this is not the direction - in apparence - this project seems to take.
Indeed, most of SNOs (myself included I must admit) use dedicated storage hardware, or even dedicated servers, to run a storage node. Sometimes there are discussions on the energy efficiency of the hardware, but with dedicated hardware we will never reach the same level of energy efficiency measures that can be implemented in a traditional data center.
I would like to better understand if people are running - successfully - a storage node on some existing hardware that would run anyway and on which you have not made any hardware upgrade dedicated to Storj.
Also, I would like to know if Storj can be compatible with the use of a simple, 1-hdd NAS, occasionally used to store docs, cat pictures and holiday movies (what ordinary people have at home basically).
If the answer to my two questions above is yes, this will start to reconcile my initial expectations with the real-world use cases of Storj.
I believe using only what you have does not work out well for 2 reasons:
The income is not worth it if you rent out only 500 GB or less to Storj
The space has to be dedicated to the node without the ability to reclaim it back instantly when need. So normally you would prefer to run a node on a dedicated hard disk rather than a joined one for normal use.
The footprint so far is minuscule. About 8000 nodes means at most 8000 computers, likely much less than that as people run many nodes per computer; also many nodes would be 24/7 online anyway for other reasons. A single aisle in an average data center has more computers.
However, per-TB it will probably always be worse than a centralized data center exactly because centralization helps keep impact low. A data center NAS will probably be a single server with 40 or more drives, not one PC per 2-3 drives, which is what I expect an average SNO have. Also it is easier for data centers to improve carbon-neutrality, again due to economics of scale available due to centralization.
More because energy is usually costly for home users, not because of environmental impact. My node is on a PC that costs me about 25 USD/year to run in electricity costs; for my nodes this means about 20% of what they earn.
The income is not worth it if you rent out only 500 GB or less to Storj
It depends which region of the world you’re in. It can become quite substantial in some areas. Also I guess that there are more and more NAS or residential gateways that have several TB worth of disk space.
The space has to be dedicated to the node without the ability to reclaim it back instantly when need. So normally you would prefer to run a node on a dedicated hard disk rather than a joined one for normal use.
For me this is a much stronger argument and I am wondering how to overcome this. Couldn’t there be a functionality enabling to use a variable amount of storage for Storj (e.g. all space left on the drive) instead of a defined disk space?
I am more thinking long-term, when hopefully Storj reaches the same scale of major cloud providers. When this milestone is reached it will be a considerable advantage to be able to perform better than traditional centralised cloud providers in terms of environmental footprint. Everyone keeps talking about the consumption of our devices, networks and data centers, the latter being regularly pointed at for their bad environmental footprint. So I don’t know exactly how, as an innovative business in this sector, you can avoid this discussion nowadays.
STORJlabs pays SNOs in STORJ, which need processing power on the blockchain, making it not environmental friendly, at least until ETH doesn’t use POW anymore I guess.
Looking only at the storage side of Tardigrade/STORJ wouldn’t make much sense imo.
But yes, payments are only once a month, so I don’t know how big the impact really is.
I think something like this could be implemented quite easily as it is just a relocation of existing pieces no repair is required. And egress would not need to be paid as it is the request of the node that pieces get moved off from it. Ingress to other nodes is not paid as well so it would be basically at no cost for Storj.
Other nodes would be happy as they receive pieces from another node. And node owner is happy as he could reclaim need space faster than before.
The only downside is that malicious SNOs could abuse such a feature to get rid of data with low egress to increase their space after that again and hope for new data with better egress.
My server was on 24/7 before i started my storage node because of the homeserver functionality it already had, so the only extra power i’m using is the extra hard drive i added because of storj. I would dare to say this goes for a bunch a sno’s? I can’t imagine everyone is running a dedicated server purely for storj? Dedicated drives, sure.
My nodes run almost entirely on hardware I already had. I just expanded the array a little. With one exception, I moved 1 HDD to an external enclosure to run a node on there. All of them are attached to my NAS which was already on 24/7. I already had a 4 bay drobo attached and online 24/7 as well. Personal storage requirements don’t grow all that fast, so I don’t mind assigning storage space semi-permanently. By my estimation running my nodes uses about 10w on top of what I already had. Can’t get much better than that. I share a total of 28TB.
So… it’s definitely possible.
I believe partial graceful exit is on the backlog, but I don’t know whether it’s being worked on yet. This would definitely help those who don’t have as much to share and don’t want to run out of space themselves. For me, so far Storj has easily paid for any expansion my node required, so I’ve just been buying more HDD’s when space ran out.
