Running a Full Node

So, I have been researching a bit with a friend on making running a full node easy through automation on (different) cloud providers. As said elsewhere, OVH is relatively cheap with $160, packet and IBM next with $360 and then Azure with $700. AWS and Google seems to be out. I am gladly corrected :slight_smile:

The big one is still really SGX because as we now learned it really needs a full Intel accepted license to work and there seems to be a lot of documentation missing. And especially automation will be hard for that one.

So I thought I’d ask: is there even an interest by the foundation for something like this to exist or is the goal rather to have enterprises run them instead of individual people? What are the goals of making the validators completely distributed and when we are done can the node just join or is there more interaction needed? And I mean that seriously and all the ways are possible.

Someone from Blockdeamon here? Will you be providing a simple service for this anytime soon?

2 Likes

Actually just mostly answered in the signal group. Running a validator needs real world trust meetings for now. More to come there but this is now.

This this can’t be automated. Good to know.

Great questions in the OP thread.

I’d be interesd in answers to your questions, too.

I don’t follow discussions on the Signal group (not time, sorry) but I’d be interesd in a digest of statements here.

So if you’d like to elaborate, please do so! :slight_smile:

Based on my current information there is no way to easily run a full node.

a) you need to get the SGX going which means an Intel License, which is doable but annoying.
b) you need a real meeting to establish the trust connection because that is how you get in the inner circle of trusted nodes.

I think they need to be more specific about this because it is ok, it is just different. Currently I would say:

  • This is not decentralised but will be run by a group of trusted parties and you will need to be sure that you trust those parties.
  • I fully believe in the privacy goal and they did a great job there so it will be for private payments.
  • still unclear is the coin distribution but I am sure they do not want to do this for personal gain. Still that is unclear because the coins belong to the foundation. Remember Ripple is currently thinking about just burnings their owned coins. And Mobilecoin is using the same base system.
2 Likes

Did not know that there was a ‘SGX inner circle’ type of thing where I need to meet someone. Is that really so (in other words: source?).

I understand SGX on a basic level, but I am not an expert by any means. Nor do I have any pracitcal experience in using SGX.
Sounds interesting though.

I think/hope (my personal speculation / worldview) the following:
In the long run, Intel SGX is not the way to go.

I hope there will be a RISC-V SGX drop-in replacement (or at least: a RISC-V-Intel-SGX-competitor) and thereby no more vendor lock-in.
Thinking in borader terms, this is the only thing that makes sense to me.

Might as well consider to work on RISCV at some point if life get’s boring :slight_smile:

At least - this is the future I am hoping for. So… this is the future I will be willing to work towards to.

Cheers

getting into the full validation system still needs a wetware meeting as the notice. Which is fine as long as it is known.

SGX I think is great, it just makes things harder. If the goal is to have e.g. the big integration companies like Signal running the validators then this is actually fine.

The SGX and need for signing by Intel does create a single point of failure though. Let’s see how we move forward with this.

1 Like

I don’t think it necessarily requires a meeting in person but the design of the consensus system in MobileCoin is very similar to DNS: webs of trust. We want anyone to be able to run a node and for node operators to be able to decide who they peer with.

If you can convince the MobileCoin node operators to peer with you, then you’re part of the validator set. Every node configures its own quorum slice.

Yes SGX makes operating a node more difficult, but it is a requirement for the trust model of this system. If we want to create what is essentially an oblivious Venmo, there’s no way I know of to build such a system other than this.

That being said, we’re always open to alternative designs. If you know of a way to build a better mousetrap, we’d love to learn about it!

Cheers,
Joshua

2 Likes

If course there are systems where people do not have to meet in person or close to it to trust each other. That is the point of a large part of crypto. And I do think you have to stop comparing yourself to it. I really still like mobilecoin and it can really succeed as an exchange of value system on different mobile applications. You had to do decisions to get to that vision and some might be different than other people would have chosen.

You need to celebrate not defend them:

SGX is pain in the butt to run and a single point of failure but it was needed to get to the privacy level you wanted to get to and long term there might be other solutions. Not now.

You want small fees and high throughput that you can guarantee and that required a way more centralised approach for the consensus system and also didn’t give you a way to incentivise node operators but you believe they will be the people that integrate the service anyway and hence that is fine. Can they stop the service? Yes. Can they get to your wallet? No. Also good enough for the use case.

This might not be an investible asset but that is not what you want anyway.

You have great strengths and I am looking Forward to use the product.

2 Likes

Technical wise it’s a great product!
@ coin distribution
Marketing wise it is a desaster!
On Dec 10, 2020 (https://twitter.com/MobileCoinTSvc ) MCTS announced:
We are excited to announce that MobileCoins are now available for purchase. MobileCoins are designed to send and receive peer-to-peer payments securely and privately. Visit the official sales channel at http://buymobilecoin.com. It’s now nearly 5 months ago and what happened?
Nice chats on The FTX Podcast - Joshua Goldbard Founder of Signal Payment Token MobileCoin - YouTube
Why is MobileCoinTSvc not distributing (selling) some of the 250 million coins?

That is simple: It’s really really really hard to fairly distribute pre-mined tokens. Ripple, who use the same system, might be burning theirs, as an example. And they want to do it legally, mobilecoin that is, which i applaud them for.

1 Like

If you have a proposal for a 100k+ node system that still has 3 second transaction clearing time, I’m all ears. To my mind, decentralized means that no single party has authoritative control of the system and by any measure MobileCoin meets that criteria.

I don’t think it’s correct to say that MobileCoin is something that any service provider can stop. You’d need collusion between the node operators to halt but the encrypted ledger data is public and anyone could reboot the network, just the same as any other cryptocurrency.

Ultimately MobileCoin will have many nodes run by many different orgs. Today there’s no org that runs a significant portion of the network. If there’s no authority that can unilaterally halt I’m not sure what we’re debating anymore.

Cheers,
Joshua

As mentioned, you made the right choice for what you want to have. But I would presume you understand that somebody that believes in Bitcoin and Ethereum with all their heart will not consider you truly distributed.

I presume intel could turn you off by removing SGX usage certification. As nobody can run a node by themselves hence people are wondering who is running nodes.

I believe in Mobilecoin. But I guarantee you:

  • Saying you are not centralized is not believed out there. Not without proof. Ripple and Stellar tried for years and failed.
  • Saying e.g. Intel can’t turn you off needs to be proven.

We are talking about crypto community, not everything is logical in arguments out there :slight_smile:

I want this to succeed. I will shut up now as I am turning in circles.

1 Like

Not an expert - not yet - but I know of no reasons against a RISC-V based SGX replacement.

Long term, I think this needn’t be an issue as long as there are people willing to make it happen.