Proposal for Node Provider Self-declaration

[Sormarler]

The point of the NNS , well besides to govern the network, is to make a potential disappearance of the foundation uneventful. The NNS is the gatekeeper. It is solely responsible for onboarding new node providers.

[Zane]

@SvenF

Second, new NP will have a substantial stake in the IC through the investment in HW Infrastructure for running their nodes. As described by Dom in his recent article (Blogpost Dom ), NPs are not required to stake ICP, but the HW investment ensures that the NP has sufficient incentive to ensure to run nodes efficiently and reliably.

I also don’t agree with this statement by Dom. A stake in ETH:

• ideally increases in value.
• is locked in the protocol.
• can be slashed as soon as the node misbehaves, causing an instant monetary loss and can be slashed to 0, causing a complete loss on the initial investment.

Hardware:

• loses value as time goes on.
• can be resold or repurposed.
• the penalty for misbehaving nodes is reducing future rewards, which ironically also reduces the incentives for that provider to behave correctly, especially once the initial investment has already been paid off by rewards, at that point the provider has no real incentive not to misbehave other than losing a revenue stream.

[NickM]

Same here, good comms, useful to be challenged in a positive manner and to freely challenge, appreciate it

[SvenF]

Thanks @JaMarco, that’s a great question! We do not measure decentralization in terms of ISPs but will take this back and discuss with the team.

[Jimpachi8]

@Zane Strongly agree with your opinion.

On the other hand, I would like to try to understand what Nick Szabo says (which Dom quoted in his article).

Small-game fallacies

[SvenF]

Dear all,

The motion proposal for the self-declaration process is live, please find it here: Proposal: 98547 - IC Dashboard and vote to adopt the proposal if you support this decentralized Self Declaration solution as part of the NP onboarding process.

Thanks all for your comments in this thread. Based on your feedback and that from the Node Providers, the process was further simplified by allowing a business registration document (if the Node Provider is a business entity) or any personal identification document (if the Node Provider is not a business entity), as long as the documents can be verified by the community. No other changes have been made compared to the draft proposal.

Some relevant feedback has been received, and the process will need to be refined further in the future, for which new proposals will be submitted. To summarize, some of the comments and suggestions for a future process are:

• How to store and share the self-declaration forms on the Internet Computer wiki, making it visible to the community as a whole.
• Storing a hash of signed self-declaration forms as part of the NNS proposal for new NP.
• Investigate whether self-declaration forms can be stored on-chain.
• Type of documents valid for the identification of a business entity.
• Leveraging existing processes like Internet Identity to improve the self-declaration process.
• Community and legal measures that could be applied when not acting in accordance with the self-declaration.
• Future decentralization of the self-declaration process.
• How to apply this process for different geographies and countries.

Please continue to share any additional thoughts and suggestions on the self-declaration process in this Forum thread, looking forward to your suggestions!

[dfisher]

Will Dfinity educate the public on how to effectively run a KYC process? How is the community to know if the self-declaration provided isn’t just fabricated by the Node Provider?

[skilesare]

ICDevs will vote accept because it is nice to have a process, but it seems that this is ultimately up to the community and voter to verify. We need a better set of institutions/groups to validate and endorse. Ideally we’d be able to follow a group of knowledgeable names neurons that are voting in admiring neurons and holding them to a standard. ICDevs does not intend to do this vetting ourselves and it would be outside our mission to do so.

[wpb]

This is another example of needing to have a way to incentivize people or organizations who are willing to perform this type of work. We are effectively asking people to perform real work as volunteers. I just don’t see how this is sustainable.

[skilesare]

We have a gap between “dfinity does this and I trust them” and “I need to take responsibility for this”.

Theory: If dfinity stopped voting on replica management then node providers would be forced to get involved in verification. Why? Because iI’d submit a replica that sent me an ICP every block.

[SvenF]

In a fully decentralised network, I agree there should be a group of people that take on the responsbility for the onboarding process including the KYC procedures. ICDevs might not have affinity or incentive to do this, but node providers would have. Node providers are currently already sharing experiences with running nodes. Managing the KYC procedures could be another joint effort. Dfinity being also a node provider itself could take the initial lead and bring the node providers together on to discuss this. What do you think?

[lastmjs]

@Zane

and more of a trustless computing platform where devs can run software and have realistic guarantees its execution won’t be tampered with and users can verify the software they’re interacting with does what they are told and both devs and node owners can’t change that without them noticing.

Is this not what the IC is? This is more along the lines of what I think the IC is and maybe should be. All the talk of basically subverting governments is possibly dangerous, shady, and reckless. I know there’s a lot of nuance and we’ve chatted before, but I just don’t feel comfortable getting on board with the idea that we can deploy any piece of software that is above the control of any legal society on Earth.

And even if we could deploy truly unstoppable software, governments could sanction it and effectively stop its mainstream use. And mainstream use is maybe the most worthy goal in my mind.

Maybe a better statement is: what you described sounds like a more practical and realistic vision for the IC.

[Zane]

Yep that’s what it is and it’d be completely fine if that’s what we were originally promised and currently being advertised, but it is not.

Watching the old presentations, some of which I’ve linked, Dom paints a very different picture:

Business & the Blockchain - Dominic Williams

The vision might have changed over the years and if so it should be reflected in the marketing approach. If the IC is more of an AWS with blockchain features than a decentralized and sovereign world computer then it doesn’t make sense to compare it with other protocols, which try to solve a whole separate set of issues, to flex its capabilities.

It’d be like saying a commercial airplane is superior to the space shuttle cause they both fly but the former is cheaper and carries more people, true but completely ignores the fact the shuttle can go to space. The crypto industry is trying to build a commercial shuttle, is it possible? Who knows, but it definitely is not ICP and should not be sold as such.

[Sormarler]

It’s a decentralized world computer just like other gen 3 smart contract platforms.

[lastmjs]

I wonder if a slightly different take is more accurate, or maybe I’m about to say the same thing in a different way.

The IC is a decentralized computer like the rest of the blockchains, but it has chosen a different set of trade-offs.

Various blockchains have their various trade-offs, but generally they all seek decentralized compute as the end goal.

[jsull9]

First, I want to applaud the individual(s) who drafted this template.

However, I am concerned that this is not/ will not withhold in any US court system. From my view, this is as legally binding as a homemade contract between two friends with no notarization. It may/ may not hold up in court systems.

There are serious repercussions/ implied liabilities that could actually cause harm. However, if this document is not notarized and legally binding then it is sadly, just a fancy motion proposal. I applaud and agree with its intentions. I just feel this should be something we actively test in court systems on our own.

For example, DFINITY should run a “test Node” and test all of these scenarios in court without any impact. Simply run a node separately and test these different things with legal teams or mediators. They obviously have legal staff retained for the NPO. I feel we should assume this will happen in due time. If/when it does we should not be leaving it to a motion proposal.

they could do something like this

1. run the node and maliciously attack it themselves through the various vulnerabilities currently.

2. “take themselves to court” instead of waiting for it to be used against them or the blockchain. They could use the legal staff to do mediation with these scenarios and test all of the options and come back with fixes to the holes in the current system.

3. Instead of having others “attack them” we could be preemptive and “attack ourselves” for the sake of ironing out these details.

4. Hire a notary to make documents such as these truly legally binding not just by the blockchain but by US court systems.

What if, we neglect these issues and “just wait and see” and because of that a court system flexes its muscles and poof? All of these make no difference because they forced boundary nodes down and US investors lose access to their neurons/ NNS dapp

Are we really leaving it up to a (well-written) motion proposal? I just don’t see it playing out in court the way people might assume or feel it should under these terms/ limitations to hold anyone truly liable for malicious attacks or interference.

[witter]

@Zane

Anonymous nodes = safety in numbers provided by statistics and game theory
Known Node = illusion of safety provided by an imperfect and cheatable system (KYC)

Cann’t agree with you more.

[witter]

@Zane

If the 2 points I’ve mentioned aren’t solved, there is nothing stopping a well funded organization from adding many nodes under different NP accounts other than the sum of money required to buy enough rigs to obtain a subnet’s control.

This is exactly what I am most worried about, they can control the nodes, do evil to gain enough value and then let the IC to die.

People break legal contracts every day regardless of the consequences.

Even in areas where the law is functioning effectively, the law only provides minimal protection, do you see how long has it been since the mtgox bankruptcy was effectively handled? Do you see the global loss of customers due to ftx being compensated because sbf was punished? We do live in a world where the government makes the rules, but I think one of the core values of blockchain is to create a trustless environment, thus preventing fraud and break legal contracts

[witter]

@Zane

t is if subnets’ node count stays the same and without automatic node rotation between subnets. The latter has already been discussed and might be eventually be implemented, on the other side a node increase big enough to make anonymous node as safe if not safer than deterministic decentralization is unlikely to happen for 2 reasons:

1. As the subnet grows in size time to finality increases too, this could be solved by implementing something out of Dfinity’s original design for the IC: randomly sized committees.

DFINITY: Scalable consensus with threshold groups - BPASE '18

2. With bigger subnets, the tokens minted to pay the node providers would increase linearly, I’m already not certain whether the IC can become deflationary in its current state, let alone if subnets had between 400-1000 nodes in them (number of nodes planned in the original design), that’s a x30-x70 increase in minted tokens per subnet, so either operational costs for devs would have to be increased or providers rewards decreased, both by a LARGE margin. That is in my opinion unfeaseble with the current protocol revenue structure.

Effective and reasonable economic incentives are an important basis for IC to be sustainable, and the issues you mention are what concerns me.
According to the current incentive method, I think it is a subsidy strategy, similar to the planned economy model, this strategy lacks the flexibility space for DAPP increase/reduction, the increase/decrease of nodes is not flexibly adjusted because of the increase/decrease of IC network’s demand for computing/storage, the nodes are just a kind of positions with specific salary, and the lack of flexible strategy for salary adjustment will lead to the increase /decrease at an unreasonable level for a long time.
At the same time, the lack of pledge mechanism makes the NPs lack of interests consistent with the IC network as a whole, and once the NPs find a way to do evil to gain more revenue, the NPs betray the whole IC network.

[witter]

@Zane

I also don’t agree with this statement by Dom. A stake in ETH:

• ideally increases in value.
• is locked in the protocol.
• can be slashed as soon as the node misbehaves, causing an instant monetary loss and can be slashed to 0, causing a complete loss on the initial investment.

Hardware:

• loses value as time goes on.
• can be resold or repurposed.
• the penalty for misbehaving nodes is reducing future rewards, which ironically also reduces the incentives for that provider to behave correctly, especially once the initial investment has already been paid off by rewards, at that point the provider has no real incentive not to misbehave other than losing a revenue stream.

The important thing is that dom is done in a way that doesn’t align the interests of the NP and the IC network, the NP is more like an employed person, and employed people leaving and joining a new company happens every day in this world (the NP can use the hardware they put in now to do provide other services, or even sell that hardware).

As a member of the IC community, I don’t think the ETH pledge is the best model, but at least with this model, the interests of the ETH POS pledgee and the ETH network are aligned, what is the IC model for aligning the interests of the NP and the IC network?