What does the "endgame" or "midgame" of Storj farmers look like?

Ideas & Suggestions
1

Correct me if I’m wrong here, because I have not researched this.

Bitcoin and other Proof-of-work coins have long passed the early stages where a simple laptop would have sufficed for mining.

This is a map of Proof-of-work cryptocoin mining locations:

image
image2173×1172 404 KB

Most coins are now mass produced by long rows of raspbery pies in storage rooms in countries with cold climates or up in the mountains, in order to achieve the lowest bang for a buck.

So now I’m wondering what Storj is going to look like.

For starters, the clearest knowledge that I have of Storj is that
HDD > SSD, because it’s cheaper and will remain cheaper for quite a while
and there’s no penalty in using them.

But I’m already struggling with NAS vs. raspberry Pi’s.
NAS servers are more expensive than Pi’s, so naturally one should buy Pi’s instead and then have them work with HDD.

Then there is only second-hand machines vs. buying brand new.
Second-hand has risks involved and they get greater every month.

Can one switch their old HDD’s for new ones and not have to start over again with the node?

And then there’s big vs. small. Are early adopters going to create giant storage facilities, because those already exist and they’re owned by Google and Amazon.
So does that mean that even when Storj becomes popular, Google and Amazon will eventually jump to using Storj and because they can buy in bulk and then shove everyone to the side.

Or will it be the other scenario, where small and more spread out is better.
But that means that an ISP technician or perhaps yet again Google/Amazon will go into every home and install a Storj computer there and then yet again, the largest companies win again.

1 Like
2

For Storj, as it is designed right now, you mostly need connectivity. Raw electricity, not really. The end-game will therefore likely be data centers located close to internet exchanges. Besides, it would already be more profitable for some existing companies to set up storage nodes as opposed to selling the disk space directly, if only Storj had enough traffic to fill up existing supply. As soon as real demand for Tardigrade services materializes, Storj will need to actively fight centralization if they don’t want to end up with a bunch of data centers hosting most of data, probably by changing the incentive structure.

1 Like
3

They can do that? If they can, then there’s hope.

4

Datacenters is highly unlikely to be a done chapter for a long time…

just like today we are trying to push more cores and ram into a cpu die and more cpu’s on to motherboards to increase their interconnectivity, this helps with processing of various tasks…
in that sense datacenters are no more different than like safes.

they have their usages, and datacenters are much the same, they have their own advantages, just as distributed storage has its own set of advantages and disadvantages.

some of the customer base will migrated from datacenters to distributed storage, just like some crypto users will be coming out of the bank infrastructure.

hdd’s may very well die out in the coming years… already we are seeing SSD pr TB prices drop quicker than HDD pr TB prices, so soon there may be no advantage to even consider using HDD’s
aside from that, the advantages of using SSD’s rather than HDD’s could also be rather significant, you are paid per egress and egress is counted on who is the fastest, so SSD storage gives a slight edge… however the internet latency is often much more so, it’s rather insignificant atleast presently.

and the there is the power saving… but again almost insignificant.

yes nodes can be migrated across hdds and platforms

the storagenode isn’t software its basically just a database and a big folder of files…
thus you can move it between anything basically so long as you can download storagenode software to run it.

don’t really see this as a problem… i’m running a 10 year old server, its fine… i trust that much more than i would trust a new RPI, not sure non HA tech is well suited for 24/7/365 operation.
ofc a few RPI’s can be built into a cluster, but that just presents more problems i guess now that i think about that idea lol.

the time for early adopters i think has come and gone, now the long haul phase has begun.
the network has been live for enterprise usage for a year, and seems fine without knowing to much of the details of that… can’t remember storjlabs having reported any major issues.

if the majority of nodes was collected in one or to limited places then there would be no point to the entire network… so i’m sure they will make sure that doesn’t happen… and it’s really not that difficult, because stuff goes down from time to time… it will always be clearly evident over time which nodes are running from the same locations and i’m sure they will take measures to ensure this will not become a real problem, besides one would think it would sort of balance itself out or atleast can be automated to do so.

generally companies of a certain size don’t really use mediators, they will just buy the entire company if they wanted to use it…

however the tardigrade / storj network is planned as being fully decentralized long term, not quite there yet afaik, so at that point it doesn’t really matter anymore… but companies like google or amazon will most likely build it themselves rather than pay another company money for work in their chosen field or whatever we want to call it.

the biggest trees in the forest casts the most shade, bigger workforce more power…
sure it feels unfair, but want can one do… argue that bigger numbers should mean less… doesn’t really work like that… :smiley:

1 Like
5

Well, it’s not baked in into the protocol, so they can do so easily. Like, let say, reducing payouts to IP blocks that host more than 1PB of data and increasing payouts to those with less data.

6

War. Start a civil war, or start a world war.
I would like to avoid help inflating such a crisis and avoid being part of team ‘lose’ when it happens.

That’s great!

If Storj can ensure payouts to remain decentralized, then I will argue on this forum for them to do so.
If they’re not gonna and/or be selective for another reason, then I know this cryptocoin isn’t for me and I’ll have to look further.

7

Long story short storj has it covered in a Way against Storage Farming, all Nodes on the same /24 subnet are treated as one. this is because it tends to be that /24 subnets are located in geographically similar areas, additionally I have 2 Nodes in DataCenter as a plan to diversify my portfolio but storj isn’t the only thing I’m hosting at the moment.

And datacenters are also limited in the count of how many subnets they have access to, I know the guys I work with have 4 available soon to have 5th from what I heard. but the maximum per physical machine is 2. They restrict it. I have 2 systems deployed as I currently use 2 of they subnets for hosting, If you are wondering I do not have any neighbors.

So storj is Currently protected. and ISP`s nor Gigantic tech wont do what you described because of limitations
Cloud storage companies legally would have issues because of European GDPR and also other issues.

and ISP`s cannot do what you described as that would have a knock on effect on the network and would have access to limited quantity of subnets.

8

Hahaha that is cute… I have a cheap vps with 5 ips from different subnets and could get up to 20 ips…

9

I’m not sure so do not quote on me but isn’t that against the user agreement

10

Funny thing, it actually isn’t. It’s against the tos that I run more nodes than I have physical cores :rofl: but not that I forward IP Adresses from a VPN.

3 Likes
11

You’re in a race with other nodes, if you run things through VPN’s and have multiple nodes on a core, you’ll likely be slower than the neighbor who is not doing those things. So you’ll lose the race more often than win it.

Storj’s network will eventually marginalize down to the hosts that can earn slightly over cost. Those that don’t earn to cover cost, will abandon unless they are enthusiasts. It’s the same with mining. But the difference here is that if you’re closer to the delivery point, you should win the race more often and earn better than the guy who is further away or sub optimal. So, who is profitable will be determined by how close you are to heavy use, and how optimal your setup is, and how cost effective that is as well.

1 Like
12

Might be true theoretically and in the future.
At the moment my ips on the vps are generating the same ingress (or even slightly more) than my home ip.

CPU requirements for a node are rather low so you can easily run 5 nodes from different subnets on one core if you have a good consumer/server CPU. Successrates will stay >98%. A pi of course won’t be able to do that but a pi doesn’t offer enough sata bandwidth and iops anyway.

But yeah, eventually the profitability will be more important. Renting a vps costs money and ingress is already very low. If we get more nodes, it will be even lower to the point where the vps costs will eat all your earnings for a year if you start from scratch. For existing users that might be different. But I only rented a Vps a few months ago… Whish I had done so over a year ago lol…

2 Likes
13

My nodes cost about $20 a year to run, with total space of 12TB.
For one or two dollars a month, who would bother? I think there has to be a base earning level per IP. For me it is probably about $20 per month where there is no interest any more.

To me this means the end game is nodes being run in regions where a few dollars is important, run on stolen/borrowed power and internet

1 Like