How to run a satellite node?

I can find docs at official website about how to run a storage node, but I can NOT find how to run a satellite node. I am curious why storj have only 4 satellite nodes. so I have the following questions:
(1) Does Storj official allow others to build and run a satellite node?
(2) If yes, where can I find docs how to run a satellite node? Want to investigate if the hardware requirements are too high or the reward is too low to stop others build a satellite.
(3) If no, why not? If so, Storj is the same as the traditional centrallized storage system, isn’t it?

1 Like

I believe since 2023 you can run your own satellite (community satellite program), but it will not be a part of Storj official network. You run your sat, and you convince clients and SNOs to join. You take the money from clients and you pay SNOs. It’s your network. If it’s rewarding or not, is entierly up to you.
It’s not like in crypto projects were you run validators or other types of sats and get a part of the reward. Here clients pay you in Fiat or whatever and you pay your storage providers and other expenses.

https://forum.storj.io/t/announcing-community-satellite-pilot-program/21666

https://forum.storj.io/t/satellite-info-address-id-blobs-folder-hex/17183

Moderators will add their input shortly, if I missunderstood something.

4 Likes

Sorry, grammar-nazi moment here.
“Fiat” is not an acronym so shouldn’t be all-caps.
The FIAT acronym is for an Italian car manufacturer. Although I guess clients could pay in automobiles :grin:

I’ll stop being an arse and crawl back under my rock now.

6 Likes

thank you. I will read them carefully, they are very helpful info.

In fact it probably is the most centralized of all cloud solutions. Over 90% of traffic is handled by just 2 satellites.

But as a SNO I don’t care, not my data…

Sounds not good.
Expect Storj can do something to improve decentralization for satellite. Because I am trying saving a few personal pictures on Storj.

Satellites are independent. So more satellites would change nothing in terms of centralization.

I personally think Storj really want to be a leader of decentralized storage system. So I suppose Storj official will do some strive to improve decetralization. Community satellite program maybe is a kind of trying, I guess. Hope Storj can do more on it, decentralization is very important to web3 storage.

They should have designed it decentralized then. :wink:

The decentralised part is the one that matters: the storage.
Having an entity that owns the satellites that keep the records and manage payments and draws proffit from the network it only makes things better.
Imagine if you have satellites working on pi3 or other underpowered harware that can’t even generate a bloom filter. Your decentralised network will collapse pretty soon because you allowed everyone to join.
Be smart and weight the pros and cons; don’t just take an idealist system as perfect for everything, if it’s working for something.
If everyone could be the midleman this just opens the doors for abusers and CIA routing all traffic. :wink:

3 Likes

What exactly is the benefit for a user from having the storage decentralized this way? Without the satellite the data might still exist but there is no way to access.

What is needed to make it save and really decentralized is a way for nodes to announce all stored segments to whatever satellite they want. But by implementing such feature storj would give up control. I guess this is not what they want…

Yes, Storage decentralization is necessary for decentralized storage, but it is not the only, because the storage is based on satellites. Storj storage system will collapse if the centralized satellites meet risks (eg. The entity that owns the satellites is controlled by the government, or occurs moral risk, etc).

Why not 10 or more entities (eg. EOS has 21 nodes)? Obviously, it is much safer than only one entity. Even if several entities occur risks simultaneously, Storj storage system can run normally.

It can be controlled by some rules. Rules are just the way our world runs. Eg. (1) setup onboarding rules for satellite runners; (2) there be a set of punishment methods to ensure the satellites run normally; (3) etc…

I think that to promote the decentralized level for satellites should not be thought as to build an idealist system.

1 Like

A good read.

1 Like

Decentralization is just part of marketing. Perhaps this will change in the future, but obviously not soon…

In the real world, the winners are the price, the political aspects of the country and the delivery time for a particular customer.

1 Like

I DON’T think we should equal ‘decentralization’ as ‘Full decentralization’ when talking about the word ‘decentralization’ (we should not think of sth with an extream way). Giving that 1K-node is ‘Full decentralization’, then 10-node or 5-node (controlled by multi entities) maybe is ‘enough decentralization’ for Storj satellite decentralization to avoid centralized risk without big speed slow down.

Storj is a distributed cloud, where all nodes are distributed across the globe, include satellites and their databases.
@snorkel is right - you may run your own satellite and attract your customers and Node Operators to use your satellite, you need to implement the payments acceptation from the customers and pay your Storage Node Operators. Perhaps you may convince Operators run nodes for free, but I’m not sure.

This is not a problem. The true problem, that there is no really fast decentralized databases. All known algorithms are requires a consensus, thus very slow and become slower and slower with the grow. No one solved this issue so far. See blogs I linked above.

Ever thought of teaming up with universities and IT-Students to work on that?

Yes, unfortunately. Seems this is needed only to distributed cloud storages…
Everyone else is happy to send money with the Blockchain.

Everyone else is happy to send transaction information (=data with a certain structure to them) with the Blockchain. FTFY :wink:

If everyone else is willing to store data on a blockchain (which is a distributed database) why not rethink if we are doing something wrong (aka holding the tool wrong or using the wrong tool for the job)? Not store actual client data (segments) on the blockchain. But audit progress for example can be completely onchain: The satellites broadcast a list of segments to audit, certain nodes get the list and fetch the segments node > node. When they are done, they add the hash result back to the blockchain where the satellites step in and see if everything matches. If not, either the auditing node is wrong or the audited node is wrong. We just solved the problem of fully decentralizing audits using a distributed blockchain. When the current audit “epoch” (=round) is done, zk-rollup the result. The blockchain’s size is still the same(ish) as it was before, like mina does.

Not saying do exactly this, just saying we should all think if we are indeed using the right tool for the job.

1 Like