There was a way to approach this suggested by lastmjs and Harrison that involved solutions such as making nodes unaware of what they’re hosting and shuffling, combined with decentralized node operation should be a hit
Those are all still very theoretical and far away in the future.
Unless a government can force a provider to shut down regardless of jurisdiction, then I don’t see how it’s possible. Let’s say a project is running on a subnet with 13 or 30 nodes and only 5 of them are in the USA. Even if the USA manage to order a data center or provider to shut down the ones in its jurisdiction, wouldn’t the subnet or applications keep working anyway? I might be totally wrong on this. I am assuming that it works.
The NNS remains a point of weakness untill it’s truly decentralized.
i didn’t say this… no one said this.
i said “reimbursed”, very clearly, which is true, the costs are reimbursed 100% over time. not sure what your issue is, I just pointed out facts and you are reacting like its a personal attack to things I didn’t even say.
-
Even if governments couldn’t force a provider to shut down in another country, they’d still have authority over those in their own. Providers have no control over the canister they host, Dfinity’s solution is shifting responsability to boundary nodes, which will be able to filter which canister they serve, but will this be enough? If the new Wikileaks will be hosted on IC, do you think the US government won’t care the leaked files are hosted on US servers if they are only served by boundary nodes in other countries? I’m not so sure.
-
If a node provider is hit with a cease and desist, he can either shut down the node and stop receiving rewards, which will mean incurring in monetary loss and lower other provider’s trust in the system, or make an NNS proposal to remove the canister, at that point voters will have to decide what’s the lesser evil: protect the canister and damage the network by slowly losing providers to cease and desists or protect the provider and get bad publicity.
-
Most countries have similar laws so even if nodes in a subnet are spread around they’d still have to comply with the country’s request eventually.
-
In the past the US has jailed people for violating US law, even tho they weren’t US citizens and never did anything illegal in the US.
In this case, no amount of decentralization will help if a government as powerful as the United States decides to go after you, right? All of these issues will inevitably arise at some point. It will be interesting to see how we respond when they do.
Not having nodes hosted in data centers would help a lot: it’s harder to send thounsads of cease and desist and there is lots of plausible deniability for the average joe, not so much for a node provider. Try seeding a torrent with a dedicated server, you’ll get a notice from the hosting in 1 month max, do it with your home connection and unless you leave in specific countries you won’t have any issues.
I agree. Running nodes residentially would definitely make it much harder for government censorship. However, that would defeat the purpose imo, which is to allow allow people built to built decentralized applications on chain that are suitable for all kinds of projects, including Enterprise level software. I don’t know if I would be able to trust a node provider with my business data if I knew he was running his nodes from his basement. ICP will likely never be able to reach the same level of decentralization as other blockchain that focus is on the backend.
Do you think running residential nodes is an option without major compromise on scalability, performance, capabilities?
Hard to say, performance will be impacted for sure, by how much nobody knows cause there haven’t been any benchmarks in such scenarios.
Obviously nodes running 10k hardware on gigabits connections will outpeform raspberries on home connections all the time. Some use cases like video streaming will be heavily impacted and could not be feasible or too expensive to be viable, but simpler use cases that require less CPU and bandwidth intensive operations, e.g DeFi, will greatly benefit from the increased decentralization.
i am curious to find out why you guys are so afraid of the government?
IC is not the dark web as far as I know. If you do everything right and within the law, there would be no problem.
Am I missing something? Thanks for clarifying if I do.
Of course, I am talking about the America’s and European’s government, the one I know most.
That is precisely the problem, who makes the law? Governments, you are entrusting someone else to have authority over what you can and can’t do, today it might not be an issue cause the government’s view is aligned with yours, but tomorrow? Who knows.
People hosting from residentual (e.g. badlands) is not at all a good solution. Akash networks realized this and backed away from their plans to focus on data center reliability.
The most important part is to actually utilize the compute power we have now, which I doubt is even coming close considering the specs of each node and how many niche websites are on the IC right now. We need way more adoption and projects. We just have to wait until the IC needs more nodes and Dfinity will start processing through all the independent node operators who want to join, myself included.
I think it was always known that Dfinity was going to start pretty centralized and work towards decentralization so this doesn’t concern me that much, especially if there are a few random independent nodes in the subnet such that these big 8 can’t collude and reach consensus easily.
I disagree, Akash is a completely different project, it aims to become the “airbnb of cloud providers” a very different concept from ICP’s, they want to make it possible for providers to rent their servers without a middleman, so obviously for such a use cases home hosted nodes don’t make sense.
For a chain that wants to be decentralized, uncensorable and resilient from attackers its a must have imo. They’d integrate pretty well with IC’s subnet philosophy, we already have different types of subnets based on the type of services they run, at the moment there are 2 types of subnets: application and system , with more coming in the near future: SNS, storage, etc…
A subnet with lower hardware requirements and higher replication factor would cover many use cases and make the network more decentralized.
Guess we now REALLY need badlands in the roadmap
This is true! I personally have been following the node hardware development, and they are actively working on creating the hardware to be accessible to average users. At first they had to create the data center boundary nodes (there was a tech. reason for this), and now that they’ve proven it works they are working towards complete decentralization by allowing the hardware/ node providing to be accessible. However, I will say I wish it would hurry up and happen because I am excited. It is happening though…
I could be wrong! I think the badlands is what users and future node hosts want. However, I have not looked into the issues mentioned above and am glad it was pointed out to look into on my own! Thanks!
I think this is ICP’s greatest vulnerability. I love the vision and the possibilities so much that it is depressing to see the design issues that have consistently corrupted crypto initiatives being replicated, from tokenomics to investment to lack of environmental planning and modelling at design stage (the very first BTC buyer warned about the CO2 emissions if BTC were to scale, in 2009!), and makes me extremely wary to invest. It makes financial sense as an early adopter, because I might be in a position to profit from a likely speculatory bubble, but I’m not yet convinced it makes ethical sense.
- People in the thread have spoken of BTC as an example of decentralisation, but that is not the case. BTC, Ether, Dogecoin and more are all, beneath the surface, hugely centralised in exactly the same way as ICP has all the hallmarks of replicating, fractally.
You can see that across all currencies, regardless of maturity, wealth is concentrated in a tiny number of people. If even 5 of those mega owners talk to each other, they have the power to shape the entire market and profit, and all the warnings of OP apply. So this pattern is not unique to ICP, nor a function of its youth: it is an intrinsic vulnerability in the tokenomics of all these coins.
Meanwhile, as far back as February, before the crash, 55% of BTC owners, or around 30 million people, were in the red. You can be certain they do not include the 0.01% (7000 people) who own 60% of BTC. Which means, across currencies, including this one, the vast, vast majority of investors bear all the risk, while also driving up the value for the teeeny percentage of whales.
- In governance, those same whales, often second wave institutional investors/speculators, can artificially shape the direction of the entire ICP to maximise profit, simply by outstaking neurons and setting them to auto-follow. At that volume, they will profit whether a recommendation makes it or not, and on both ends. The recent spam exploit is crude and peanut like by comparison. Once the big funds step in, which is when they consider ICP has reached the right scale for large scale exploitation, their exploit will not be blatant and obvious. It will be systemic and subtle. Retail investors will see value spike, and be super happy, with gold rushes powered by belief in the vision that is now promising scale, as well as the profits. But all that rising value, will still end up with the 0.01% of whales.
I see zero guardrails to prevent this consistent pattern, and the spam episode illustrates that good intentions or social norms are not enough. If the model incentivises speculation, speculation will happen.
- Finally, I worry about the investment pattern. The VCs putting millions into Dfinity are not doing it from love of humanity or a vision of a future internet. That’s for the retailers whose belief they tend to exploit. But ask any venture investor (apart from true impact VCs): they are in it for the exit. And what is the VC exit pattern across crypto? A new blockchain venture has a compelling concept and a bit of traction. Something people want to own and want to believe in and scalable. The offerings and projections are both grandiose and unrealistic to impossible (compare 2021 Dfinity projections with actual scale). You’d think VCs would balk, or push back. But you forget that it’s all about, and only about the exit. The moment they put millions into the overambitious venture, that venture suddenly gains social proof for retailers, who believe in the promises, forge personal identities and peer groups and communities around the storytelling, typically combining certainty of societal emancipation and exponential wealth, and the venture takes off. The VCs stay until they judge the bubble is nearing its peak, but before the flaws crack the surface. They exit then, and so do the founders, making, as centralised owners of most of the concentrated wealth, a HUGE profit. And then, again and again and again, the venture crashes and burns, taking with it that 70-80% of investors that hold 0.01% of the value. Look at how many crypto ventures have crashed shortly after exit, to see the pattern.
As I said, I LOVE the concept, the potential, the storytelling, the rigour and the vision of ICP. I think what Dfinity have achieved is groundbreaking, potentially significant, and in so many ways, beautiful. So much I don’t want to see the red flags. And that resistance in myself, is my biggest warning of danger.
Am I surprised that the OPs findings about distribution are not common knowledge or P1s? I have to admit I’m not.
Would I be surprised to find that stake is similarly concentrated outside Dfinity? I would be surprised if it wasn’t.
Would I be surprised if looking at the original projections I were to find a HUGE distance between promise and actuality? I would not.
Would I be shocked to discover recently or soon, some new players consolidating a significant percentage of all value? Not really.
Finally and above all: would I like to be offered evidence that proves me wrong?
With all my heart.
I can’t find the article but dfinity has lots of onboarding for node providers going on every single month. It’s more decentralized every day and they are doing R&D to make it a dao type thing. These things take time
Yeah, I found the Dfinity equivalent
You can see the pattern replicated. 50,000+ people owned 0.8% less than 1000 people owned 75% of the value, the remaining 24% in Dfinity’s hands. ))Dfinity is of course not for profit, but Dfinity team members have an 18% share, so ultimately do receive dividends of a sort.
Now, that was at Genesis. If I am wrong about replicated patterns, which I dearly wish to be, the data now should have improved the distribution since Genesis. However, if I was a betting man, I would bet that the distribution has likely seen even greater concentration.
I emphasise that none of this is to cuestion the motives of Dfinity or the sincerity behind the ICP vision. And VCs are pretty explicit about their profit motive, so no deception there either.
Rather I am calling attention to a set of design antipatterns that foretell a worrying conclusion.
Without knowing any private information, I would bet the concentration of ICP has greatly decreased because of the selling pressure created by taxes… but I have no way to truly know.
Just speaking as one person: every month I (a DFINITY employee) vest on my 4-year schedule, I pay taxes. That means a sizable % of my tokens have to be sold just to pay for taxes every month. Every time i stake ICP, i sell a portion of it to pay for taxes. I am sure others (if not all) have to do the same. Just taking the team for example, if you assume a global simplistic average of 33%, that means 33% of the initial 18% of ICP (allocated to team members) at Genesis will be sold purely for taxes (unless people pay out-of-pocket which is rare in my experience). That’s 6% of Genesis ICP alone.
That being said, some people merge maturity (which compounds their ICP) and others may live in lower-tax jurisdictions so it is possible that some entities have accumulated faster than others. I cannot honestly rule that out so this is my hunch.