Is there any benefit in Storj Utilizing ICP Tech?

I typically always start off with a disclaimer as I am not that familiar with the whole data storage tech or even how Storj Satellites completely work. However, I did want to raise this question.

Internet computer protocol (ICP) claims to be like this decentralized cloud where users use dedicated hardware to store data. Is there any benefit from Storj utilizing ICP technology for their Satellites ?

There is a thread that a frequently go back to titled “Single point of Failure if a Satellite is down?” and one of the Storj team members stated “Satellites should be run with industry-leading techniques for high availability and so on like other cloud services.”

ICP seems to be (from what I saw) cheaper, more decentralized than something like AWS, so I guess my overall question would be is there any use case or benefit in storj utilizing the tech that ICP provides?

1 Like

How about security? If anyone can run a ICP node how do they prevent code manipulation? Or just reading customer access grants from memory?

1 Like

That could be a good point, in general I am not 100% familiar with their architecture so I can’t really even say how secure they are or aren’t.

Isn’t ICP another generic smart-contract platform competing with Ethereum? (or is this not the ICP you’re thinking of?). If so, it doesn’t offer anything the satellites would benefit from.

1 Like

I thought the decentralized cloud aspect of it could somehow be used. Enterprise | Internet Computer

That’s some pretty good marketing: ICP is a ‘decentralized cloud’ the same way Ethereum is.

They’re saying because you could put application logic in a smart-contract… and the data it uses could also be added to their blockchain… and smart-contracts can execute in a decentralized fashion (on any node)… that it’s the equivalent of generic cloud computing. But that’s nothing like renting vanilla VM capacity from AWS.

ICP looks like a smart-contract blockchain hoping to attract ETH devs.

1 Like

Well if that’s the case then yes that is some good marketing because even in the speeches that they make they say they are different from Ethereum and other blockchains.

For example, in this video (really only the first 1:38 is the speech I was referring to).

Yes, they’ve got quite the marketing budget. ICP is a rebrand/relaunch of Dfinity… who has been trying to get people to care about their tech since 2018-2019. They want people to build apps to run on their platform… and buy their coin to pay to do so.

They certainly offer more raw compute and storage than the Eth network. Maybe they’ll be successful!


No, anyone cannot simply run an ICP node. It is not open in that way, currently there are no additional nodes allowed to be added. Approval is strictly limited to those properly vetted.


Based on the video below it also seems to be extremely expensive to run a node machine. I had read your bio and was wondering if you had the knowledge to verify/agree or deny to what Roxor was stating above about ICP.


Dfinity is not looking for node operators, the specs call for very large machines run only in commercial data centers around the world. The ICP platform looks very compelling as a way to run apps. From what I studied it is fairly easy to develop on their platform and it is extremely secure and affordable for end users. They have limited amounts of storage, but enough for their current targets. I am very positive ab out their platform for both application developers and end users. I plan to attend some of their events next week in Brussels to learn more.

Yeah… to prevent the provider rewards from being diluted any more (since the network already has way too much capacity for the amount of paid usage) they’ve been capped for about 6 months. But they could open it up again.

1 Like

The current problem for decentralization not in the distributed compute itself, but more about a decentralized secure and fast database (we again can select only two…), where even a Byzantine node can participate. Nobody invented this so far.


When SNOs discuss Satellite related topics those SNOs make an impression being not SNOs or at least not just SNOs :sweat_smile:
However I’m just a SNO and not affiliated with Storj and shouldn’t probably speak to this topic. But I still can make same analysis and give some thoughts. Please take it as just mussings and not as something to rely upon.
So conceptually yes, ICP looks like something that could run satellites. But the devil is in details and here are some considerations that immediately come to mind:

  1. I guess Storj satellite software is written in Go. It doesn’t have to be but given that the node software is Go I’d expect the rest of the software is on the same stack.
    • Quick look at the ICP supported languages shows that there is no Go support (at least now) and Rust is the equivalent they support.
    • However they look compiling to WASM so may be adding Go support is not too difficult.
  1. Storj uses QUIC protocol, not sure if it might cause issues on ICP.
  2. Based on the previous comment above I guess most computation of satellites are offloaded to the DB. Sharding is usually used to set up a distributed DB but performance still can be an issue in such setup.

So overall migrating satellites to ICP even if possible most likely will be too expensive and will not happen any time soon.

1 Like

It also doesn’t make any sense until they can provide the similar to a private network security… Thus, no improvements so far.
I would say that existing Fog computing solutions are more capable right now, but they have the same problem - any operator of the node can get an access to the data of the running processes on their node. So, you cannot run any process which operates with a non public or unencrypted data…

1 Like