There might be some misunderstanding, but that’s the life of a text-based forum. Please assume good intentions. For many of us here English isn’t the first language, tone and nuance can get lost in translation.
That said, people here are right—running a community satellite is not a use case that Storj software is prepared for now, even if it was developed with this intention. There’s only so much capacity Storj Inc. can dedicate to developing software, and this point has never been a priority. They’re trying to run a profitable company so that there is more capacity for this kind of endeavor in future.
As such, if you want to run a satellite, this will by necessity be an independent effort, with goals and limitations different from what Storj Inc. encounters. You will be an independent actor, doing all the operations, maintenance, tuning, marketing, selling, customer support, attracting storage node community or maintaining your own nodes, etc. yourself. There apparently were attempts to do so in the past, none of them truly successful. This is difficult, requires a lot of time and inside-out knowledge of the codebase. People who have this kind of knowledge don’t really hang out on this forum… they’re busy elsewhere.
You might be able to sign a contract with Storj Inc. to get at least some of that—they do offer professional services. Storj Inc. helps maintaining private satellites for some of their customers. But as paid, private consultation services, this kind of support just doesn’t happen here on the forum.
for us it does not matter, but i guess as long the trust issue for the satellite operator is there their own idea of a community run network of satellites will never get off the ground
i cant see how i can hack the data provided i just host a tamper proof black box i dont have access to anyway… there would be ssh access for storj to access it, but if i just as much as try to login once they could remove it from the network
the same if its a machine i build and then hand over the control to them
but they could be out there, but then we would never know as there would be no need to disclose it or even make it possible to sniff arround to find out. but personally i would not bother to find out as it would not help me or anyone else
I didn’t understand you wanted to build a satellite for Storj-the-company to use. That will never happen: they’re a private firm that needs control over their service - the risks of trusting some-random-dude-running-a-satellite-in-his-basement will always overwhelm any technical problems it solves. A large part of their special-sauce is how well their satellites are run.
However, the community could certainly use a second community satellite. I think even without a token for a “payout”… it could run well with a ratio system: like for every 3MB a SNO has as used-space: they can use up to 1MB on your network. Everyone using it would know it could disappear tomorrow (if you decided to stop running your satellite).
You could decide it’s costing you too much (power, space, dealing with heat or loud fans, or hardware failure+replacement). Or you pass away. Or your house burns down. Or you move and your new Internet connection is only over Starlink or something.
But everyone using your satellite (as a cheaper-but-way-less-durable option) would understand that. It would still be valuable.
the house i got is 400 meters from the west coast in denmark and the avg. sales time is 10 years, at least i can give people a warning. But honest the plan was to live here until i move in to a pine box
Eh, think reddit vs. lemmy, somehow lemmy got a nontrivial userbase.
Why would people create accounts on your satellite vs. Storj-provided satellites? What’s your value proposition here?
I can only guess, but I would suspect data locality. If you already have hardware in a data center, setting up a local satellite accessing local workers makes a lot of sense—I recall some Storjling saying it’s better than Ceph.
Of course, but Storj is an owner of this cluster and manage it, and even then only a small amount of personnel has access to the production database in a write mode.
You, as an owner of a part of the satellite, will have this access and can drop the data or disrupt the service.
Perhaps it can be done for the certified DCs with a contract, NDA, SLA, big fines, etc., but unlikely for individuals.
I believe, the only service, which could be outsourced is a repair worker, it doesn’t have a direct access to the database, it contacts only the satellite API, it’s very similar to uplink, but doesn’t decrypt the data (we doesn’t have an encryption phrase anyway) and just downloads minimum needed pieces and re-calculates all missing, then uploads them to the new nodes.
This is related only to Storj’s satellites. If you would run your own satellite - it will be your own Storj network, separated from the Storj production. So, you may set your rules and prices and the way, how it can be profitable.
@Toyoo is right, there are private satellite operators, and many use cases are to replace CEPH, because Storj protocol is more efficient and robust without heavy investments, because the storage can be on almost any hardware which they may have and the protocol is robust enough to work even in a Byzantine network with a minimum monitoring or even without it.
Of course, it’s better to use Storj Global, but some have a mandatory requirements when the data must be inside the perimeter, even if it’s end-to-end encrypted.
I have no idea. Increasing a number of repair workers will quickly overload the database, so the usefulness of this offload is questionable. They should be orchestrated together.
so for a “dummie” like me that would mean that the number of repair nodes needs to match the number of satellites and the repair node needs to work for all the satelittles
Each satellite have an own swarm of repairers and auditors. The number depends on the length of the repair queue and the database cluster capacity (you do not want to overload a database, but scaling of the database cluster is costly, so you need a balance).
I think of some kind of orchestrator and worker nodes, which can be joined to the cluster. The orchestrator will automatically deploy the required number of repairers and auditors as needed.
It could be Nomad, k3s, docker swarm, etc. But first it need to be secured that even the operator cannot influence to the running processes/containers.
