Ethereum Foundation DevCon 6 - Layer 1 Protocol Track, 15 Nov 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Social Slashing by Eric Wall | Devcon Bogotá

Description

Visit the https://archive.devcon.org/ to gain access to the entire library of Devcon talks with the ease of filtering, playlists, personalized suggestions, decentralized access on Swarm, IPFS and more.
https://archive.devcon.org/archive/watch/6/social-slashing/

Censorship resistance is not a property of the protocol, it is a property of the community. The Ethereum community controls one of the most powerful tools in the crypto universe to combat censorship—but no one currently knows exactly how to use it. Swing by this talk to prepare yourself. What's at stake? Everything.

Speaker(s): Eric Wall
Skill level: Beginner
Track: Layer 1 Protocol
Keywords: Slashing,Censorship resistance,Governance

Follow us: https://twitter.com/efdevcon, https://twitter.com/ethereum
Learn more about devcon: https://www.devcon.org/
Learn more about ethereum: https://ethereum.org/

Devcon is the Ethereum conference for developers, researchers, thinkers, and makers.
Devcon 6 was held in Bogotá, Colombia on Oct 11 - 14, 2022.
Devcon is organized and presented by the Ethereum Foundation, with the support of our sponsors. To find out more, please visit https://ethereum.foundation/

A

A

I'm pretty excited to hear his talk. um You know first heard of Eric when he was the alt calling Slayer in 2017 by taking apart uh a lot of the outlandish claims made by some altcoins um and I am pretty excited about social, slashing as a feature that allows us to help keep this space decentralized and Eric's really been a leader on this. So I'm really excited to welcome him to the stage. Welcome Eric.

B

B

If we allow censorship of user transactions on the network, then we basically failed them. This is the hill that I'm willing to die on.

B

If we start allowing users to be censored on ethereum, then this whole thing doesn't make sense and I will be leaving the ecosystem and I think there are a lot of people that think the same thing: powerful words by ethereum developer, Marius, Van, Der, Vaden,.

B

By the way, do you know that 50 of transactions, actually 50 of the blocks in ethereum right now, are being censored? They're of fat compliant.

B

Who here in this audience, recognizes this image, can I say a raise of hands?

B

What if I say the name Shaolin fry? Does that ring a bell? Okay, there are very few hands.

B

This is one of the most important symbols in bitcoin's history, and it tells a story that I think that every crypto cryptocurrency person should know.

B

So let me tell you a little story, so it's winter in 2017, the end of February Bitcoin is trading at around one thousand dollars and for the first time in bitcoin's history, the blocks are full. The mempool is full. There are rumors that there's transaction totaling, 900 000 Bitcoin, stuck in the transaction queue.

B

Now there is a an upgrade there's: a code, that's dormant in the code base that has that includes a a solution to this problem. Segregated witness has been deployed on the network, but it hasn't been activated.

B

So this is a soft Fork style upgrade and it's configured to activate after 95 of the Bitcoin mining hash rate signals their support for the upgrade.

B

It's it's the same upgrade process that the last three soft Fork has used for the last couple of years, but this time the upgrade isn't getting support. We've been stuck in this deadlock situation and we're not getting anywhere.

B

We know that, with 51 hash rate, we can break the deadlock that the miners can do it if they want to, um but we're not even close to 51 percent, so we're in a Deadlock, then out of nowhere comes this unknown, relatively unknown developer and posts an email on the Bitcoin developer mailing list, where he explains that you know this signaling methodology that we're using it's a misinterpretation. The software rules are actually always enforced by the notes, not the miners.

B

So what what this felt like by the way? This is called the user activated software. This is an important phrase to to remember so. What this felt like is basically like. Imagine that you're a baby and you're drowning in a very shallow pool, and then your mother reaches in her arm and teaches you that actually you're, not drowning. You can just stand up. You just use your legs. Your head will be above the above the water, so we call it the um the Independence Day in Bitcoin, it's the um it's.

B

The they were bitcoiners took a front seat in in the car, that's driving the upgrades, and you can argue that bitcoiners weren't really Sovereign users when they believed that they were at the behest of miners, will the will of miners to determine which upgrades make it into the protocol.

B

So this is really important, uh really important event in bitcoin's history, and this is a bit cliche this image, but really what I think happened is that we create these structures, these governance structures in blockchains, where we believe that those governance structures that we create are the only ways that we can change the network, but many times they're, just they're out of convenience. There's the only thing that is in control is the market. You can create new rules and if you get the market to follow you, then you can set new rules.

B

You don't have to abide by these structures that were created for ourselves, so we did this in Bitcoin. We were successful at it.

B

So the the uh one thing to recall, though here, is that, right before this software got activated by the users set thing and activation date in their node software, the miners actually fell in line. So the miners did what we wanted them to, because we forced them forced their hand by showing them that if they don't abide, then then we'll just implement the software for ourselves and they have to create blocks that abide by those validity rules.

B

So the actual the actual outcome Matrix and the situation was rather complex. I don't want to brush over user activated softwares and say that they're, like super simple thing, they're actually quite complex, and you can run into scenarios where there's chain splits and there's replay attacks between those Forks. So it's something that you've got to implement with care and you've got to be strategic about it, but the overall like takeaway. The conclusion from this is that block producers are merely Servants of the protocol and we don't ultimately have to abide by how they act.

B

So this story has this sort of um Cinderella quality or like a David and Goliath quality to it, where the little man stands up against the big corporations um and it's something that really galvanized the Bitcoin community and I. Think that, at this time that we see this type of galvanization happening also in the ethereum community, because the ethereum protocol is facing a threat that is perhaps much bigger than the one that Bitcoin faced. So let's go back in history and talk about the three greatest threats.

B

In my opinion that has faced the ethereum protocol, so June 2016, the Dao hack, this is was one. This was when uh ethereum's immutability came at risk and then September the same year, the Shanghai DDOS attacks. This is when ethereum's availability, the uptime of ethereum, came under stress, and then we have just the two months ago in August I, don't know exactly what to call this event, but I've decided to call it the uh the ofak. Oh.

B

And this is when the neutrality of ethereum came at risk.

B

Now, just a quick recap: you remember this. Of course, um in August the treasury Department puts contract addresses to tornado cash on the ofak sanctions list and on the 10th of August Alexander gets jailed, and what happened then was that this this Avalanche of compliance that started to threat the ethereum network, where we would see front-ends um implementing wallet, screening, which is fine, it's just the front, ends right the if the front ends decide to sensor users.

B

That's that's okay right, but uh so we saw, for example, Anthony uh sasano and Justin Sun by the way shout out to uh shout out to David Hoffman who's actually suing the treasury Department yeah.

B

He was also one of those guys that got his well. His wallet dusted and he's also was also unable to use Ave because of that, but it didn't stop there, like. We found out a couple days later that flashbots were censoring transactions. I had been doing it for quite some time. Also.

B

We now feel like the the news around our neck is getting Tighter and Tighter. Now we even see censorship happening at the Block building level and then the sort of death. Now the the the worst possible thing is when we see that even the base layer validator so the block producers.

B

So this is ethermind the I think it's the largest. It was the largest mining pool in ethereum.

B

Even they refused to implement uh sorry, they refused to include uh tornado cash transactions because their their interpretation of of the ofac law- and this is the worst possible thing that can happen right when our validators are afraid to include transactions, because the US government doesn't want them to, and amidst this we had this now Infamous tweet by Brian Armstrong, where so coinbase was, at this point in time, uh going to be the largest single largest validator on the ethereum network after the merge- and he puts out this tweet where it says.

B

Well, hopefully, it's obvious but we're going to follow the law, which is fine, but it's not super great to have your largest validator, basically saying that we're kind of feels like they also feel like they might may have to to comply with ofac.

B

So with the merge we sort of went from this situation where we had hash rate from all over the globe connecting to these mining pools and now, after the merge, we have this situation where the largest with the largest block producers, are single companies in in regulated jurisdictions like coinbase, Kraken and binance, and if you look at how many logs they make for an Epoch, each Epoch has 32 blocks.

B

Coinbase makes six minus makes one Kraken makes three: that's 10, 10 11 in total, that's actually enough to create serious damage uh and censor the the ethereum network if they, if they wanted to so I saw. The I saw these images looking at the um uh the the the live stream.

B

For this event, where two paths were being described, and we have proof of work here, on the left hand, side censorship centralization, but it's I mean I'll, be the first one to admit: I, don't know how you guys feel, but the the transition to proof of stake did bring I mean it did bring us closer to having the neutrality of the platform endangered.

B

And I've I used to think that, like crypto ethereum Bitcoin, it's this amazing movie that we get to watch from a front row seat. But at this moment I felt like I wanted I wanted to scream at the screen, because we always used to say like this is a this is a blog post by Vitality from years ago, we used to say that the um harder to harder to detect attacks like censorship, a coalition centering.

B

If that happens, we can coordinate a minority user activated soft work which in which the attackers funds are once again largely destroyed. So this is always. This was always the plan that we would do this in this type of situation.

B

um Patrick McCory has this beautiful imagery where he describes. What vitalik is really getting at is that in the ethereum protocol you can selectively Target specific validators and punish them individually, based on their actions, and we do have a track record of what they're doing right. So in proof of stake, you can there's the the actions that they do are actually attributable because they got to put their signatures on specific blocks. When they're testing, you don't have that quality improve of work, so it so in proof of stick. You actually have a superpower.

B

We have a mechanism that allows us to selectively Target specific validators, and if we can detect that they are engaging in a censorship attack on a network, we can coordinate a soft work that punishes punishes those specific validators individually, and we don't have to blow up the entire network, which we have to do in proof of work. If you want to change the work function, for example, I got to mention this I feel like I have to talk about this. There's you've seen I.

B

Think most of you probably seen this graphic where 50 of the blocks in ethereum are currently of fat compliant.

B

Is this a problem? Yeah I think it's a big problem. So what should we do like? Should we slash every single validator who's registered with an ofak compliant Mev boost relay I feel like we have to separate between between two separate issues here. So one thing- that's very important to remember is what is the reason that flashbots exists, so there's censorship in the network can happen in two different levels or actually three there's more levels, but the two main ones is in the block construction process and then also at the validator level.

B

The current censorship that we're seeing in the network is in this red box here in the Mev boost portion. So you have to remember what was the? What was the reason that flashbots exist? Well, it's because what? If in ethereum there's only one validator, that's really really good at extracting Meb there's only one validator, that's really really good at extracting MAV. They are going to be able to provide much higher, staking yields than anyone else and then you're gonna. So let's imagine that's coinbase, for example. Let's imagine that coinbase is really good at this.

B

Well, then you get a much higher yield if you delegate your stake to coinbase versus, for example, staking on your own. This is why flash boss exists. It's because they're allowing any validator to mine uh to make blocks as if they were flashbots, if as if they had access to flashbots Mev extraction techniques. So what flashbots is doing is allowing validators to not have to redelegate their stake to someone else, but still be able to mine these very profitable blocks, so they are preserving a validator level decentralization in the system.

B

Now it's come at an awful cost right. It's come at the cost that they are now in fact, compliant and they're. Building these uh blocks that everyone's using they're relaying these blocks- and this is a terrible thing, but it at least we have the at least. We have the opportunity here that validators can use other relays and there are technical ways that we can uh so basically before the merge flashbots only used to create a portion of the blocks.

B

So we can go back to such a model. We can use PBS, we can use inclusion lists, we can use partial block auctions and we can go back to a model. There are Technical Solutions to this specific problem uh which we're in now, where flashbots are creating these entire blocks and the flashbots is centering centering that that also then applies to the validator, but at least we're in a situation where validators can use other relays and we can Implement these Technical Solutions that force uh Mev Builders to include certain transactions.

B

So there's a solution to that, uh whereas if all the stake was centralized in coinbase hands and they started centering. Well, that's a much worse problem right.

B

So this is why I'm not going out on Twitter and saying, let's slash, all the all the validators that's using flashbacks, but what I think we got to do. Is we gotta explore these Technical Solutions to reduce the impact that the Mev relays and the miv builders has there's a there's a way for us to dig ourselves out of the hole that we're currently in.

B

So, let's talk about the the the more pernicious problem here, which is what, if the validators, so even if even if we um uh have a bunch of different relays that all create very profitable blocks and the validators validators can choose between those blocks. But what, if the validators themselves, they want to censor the network. They think that it's their mission to make sure that ethereum is of fat compliant. So what you have to recognize here is that there's two different ways that the validator can think about this and there's two different forms of censorship.

B

So the first one is that you basically just filter transactions that the transactions that you include, you make sure that no offic compliance get into the chain in this situation, even if 99.9 of the validators do this as long as one validator in the system does not do this, that transaction still gets in.

B

So this is bad I mean fifty percent of the blocks in ethereum being offer compliant is not a good thing, and we have to monitor that monitor that, and maybe we got to do something serious if it reaches 90 percent.

B

But it's there's one attack that is much much worse, which I call a censorship attack, which is where validators refuse to attest to blocks other validators blocks if they include such transactions. So this is a this would result in a system-wide in system-wide censorship. If validators decided to do this and I think this is the one that we should treat equally as if it was a double spend on the network.

B

If we see anyone doing rework style censorship in the network that results in system-wide censorship and that's what we really really gotta charge our uh laser laser-eyed sharks for right. This is this is the most pernicious attack. I still, don't think that it's a great thing that 50 of the blocks and it's it's a rising figure right two weeks ago it was twenty percent. Thirty percent. Now it's 50.

B

So we have to monitor that and we can't let that number go too far, but I'm sort of willing to um like maybe there's different punishments for these two attacks. Like the first one, that's more benign in nature. Maybe you can uh create a hard Fork that only allows those validators to unstake their ether and basically exit the network and it'll take them a couple of weeks. They'll take them a couple of months, it's the red one here, the latter one, the second one.

B

That is the one that we should treat equally as if it was double Spin, and by the way. Why I'm saying this when I'm, when I'm saying this I recognize that you know I'm arbitrarily coming up with rules for slashing in ethereum, and this shouldn't be something that's arbitrary. This is something that should be a policy that we all agree on, but we have to start somewhere. We have to come up with some sort of policy.

B

I think that validators want to know like what is actually a slashable offense in this network, a slashable offense in this network. So what we can do now is we can start to formulate the policy we need to get together as a community. We need to see how we feel so we can understand like what's what different routes are most likely to get system-wide consensus, so it doesn't make. So much sense for me to say this is what we should do in this situation.

B

This is what we should do in that situation, but the overarching element is that we need to come up with some sort of policy so that validators understand what are the different punishments. So, for example, uh when when uh coinbase or any other validator in the system thinks about double spending, for example, they don't need to think about okay. What happens if I doubles, but they know they know the punishment and that's something that now we.

B

This requires a social element, but we need to give them sort of similar deterrents so that they know what the punishment is.

B

Now we have bought ourselves some time and I think it's. um So this is the chief legal officer of coinbase that says that their current interpretation of the ofak law is that there's nothing that compels them to apply any form of censorship on the base layer. This is good, but I, don't think that it should be up to the interpretation of a single company or a couple of companies whether or not we have censorship in the network.

B

We need to make the decision Matrix for them look similar to what it does in the double spending case like there is no choice here like if you double spend you lose your stake, you're no longer on the network, that's how it needs to that's, that's how it needs to be for censorship. Also.

B

So where are we currently with this? So in August I made a poll where I basically asked the ethereum community to the extent that I could? So what do we do in the situation where you know ethereum is basically being censored. Do we allow that to happen, or do we slash? We burn the stakes of the validators that are making this happen on the ethereum network and it seems like currently the um sort of you know.

B

The Twitter pulse is not a great thing to to assess this, which is why I actually think that I want to be able to speak with you guys.

B

I want to be able to meet you in this conference here and ask you: what do you think that people think about this issue and you and you're going to tell me well all of my friends think that things that we should slash System, Watch censorship, attacks, real Styles and ship attacks, but we maybe should come to think we maybe should do something different if it's just you know, transaction filtering. So this is a discussion that needs to happen in ethereum um so that we can get a better grasp of what what is actually.

B

What is the consensus view on this, so um Ryan Sean Adams voted in this poll and he gave us this great logo for this result. So the the um the uh answer here to to slash the validators got the symbol: the Black Flag. Instead of waving a white flag, you would wave a Black Flag, which means, if there's censorship in ethereum, we slash those.

B

Thank you. So what is the Black Flag movement? The Black Flag movement, is sort of a just a movement of people that declare that they would be willing to go as far as to support a manual Fork of the protocol that removes the stake of validators that engages in attempts to achieve system-wide censorship.

B

Now a lot of people get iffy around this, like oh, like I, don't want to slash coinbase because they have retail users, it's ultimately their balances that will get slashed, but if you think about it, when, if coinbase double spends, they get slashed. So there's not that they're like they shouldn't, really be an ethical question. We slash them when they double spend. We slash them when they censor, like that's, not a controversial statement.

B

Of course it's more if in this situation, because we're doing it in a more arbitrarily way, we don't have uh codified rules that execute this in an automated fashion. So we need to be a bit more humble about it sure, but the ethical element to it should be sort of clear that we don't allow double spends. We don't allow censorship, so this is controversial.

B

Yeah for sure- and this is one of the reasons that we can't expect I- don't think that you can sort of rely on your traditional leaders in ethereum to be the primary spokespersons for this, because imagine metallic saying you know we're going to slash this validation, we're going to slash that validator, then we're just back to you. Know everyone criticizing us and saying the overlords of ethereum decides who gets to keep their stake or not like we don't. So this has to be in its nature.

B

It was the same thing in Bitcoin the Bitcoin core developers did not actually support the user activated soft Fork. They took a back seat. uh They took they took a back row seat in this discussion and allowed it to be a Community Driven initiative. So it's really us. It really comes down to us to push for for this type of social, slashing Fork. If that happens, because we cannot rely on our traditional leaders to do that because it just needs to be Grassroots in order for it to be authentic,.

B

Now, if you're still like on the fence about this, like one thing that I want you to remember, is that not even Swift, not even Swift applies o-fac regulations on the messages that are passed in their in their in their on their Network, because they recognize that so Swift is a cross-jurisdictional messaging system for banks.

B

If you, if you have to abide by every single jurisdictions, ofac laws and regulations, you cannot have across jurisdictional uh messaging layer. So this is a you just can't you. You can't comply with every single jurisdiction at the same time, which, which is why you can't have a network level of censorship for a global system. It just doesn't work so not even Swift. Does this and I don't think that ethereum should be any worse than Swift.

B

um There's just one last thing that I want to say here about this: this image here um it's this image uh was generated by an by by stable diffusion, so there's gpus that has rendered this image and those gpus are actually repurposed ethereum miners, so those ethereum miners that use the secure ethereum are now actually they're. Actually in this room with us.

B

Helping me tell this story so they're still with us, oh by the way I just have to get to give a shout out to Nick Carter, who helped me create this image so he's not on Twitter anymore, but he's still still also with us.

B

um Actually that was the end of my talk, so thank you very much.

B

I think we have some time for questions.

C

To stay by my side as I stumbled through the black thanks uh for the great talk, Eric really appreciated that. uh What do you think the risk is today for people.

D

To ignore the ofac guidance I mean they could tell the isps to cut off our interconnect. Internet connection. um Have you thought, through other possibilities of what the governments may try to do? If we don't comply with the request.

B

Another topic um we've seen: we've seen that coinbase doesn't see it as their role to do anything on the validator level uh to comply with ofac, so they don't currently interpret the law of the ofac law that they have to apply any network level censorship. So I think you know it's a little bit weird. That flashbots thinks that they have to do it.

B

uh I can't really speak to that on the application side, uh it gets more iffy I think that for applications, I think that, like D5 applications, maybe should not have like they shouldn't be like Incorporated in the US, maybe if they want to, but actually there's another way that you can like. If, as long as you allow the smart contracts to run an ethereum, anyone can build a front-end to that right.

B

So that's a way that you could sort of get around that issue of of you can have D5 working and then you can have other entities created from the front end Studio, smart contracts, but I, don't have a great answer to your question because it's not my area of expertise. Thank.

D

E

um Hi Eric, first of all, let me Chrome garage to lead you for the great job that you did uh debunking that hex cam I. Remember that video, the Richer hard interview, amazing I'm, a professor myself here in Colombia and at the main University and it's uh it's called the University of national Colombia, so I do a job that is similar but trying to debunk scams. People lost scams here in Colombia, so hard, sometimes, but anyway, my question is um I.

E

Don't know if you remember a few months ago that some transactions were censored for Venezuelan users through metamask and that's before the you know. The proof of stake. Ethereum system so is is the same as as you as you mentioned in your presentation.

B

Yes, I think that I recall this also, and so my stance, there is basically so private companies what type of censorship that they apply to their business is really up to them as long as it doesn't happen on the network level, because you can still use another wallet like you can import your seed from export your seed from metamask into another wallet and access those funds and move them around even if you're, a Venezuelan everything that happens on the front ends at the application layer, especially things that are like off chain is, is like it's bad.

B

That companies do that and we should try to get them to adopt a more open approach But. Ultimately, it's like Network level censorship. That's really the thing that we need to need to focus on, because someone can always build another wallet. Someone can always build another metamask.

B

So that's the only answer that I have to that question. Okay,.

E

A

Last question: real, quick.

F

Thank you Eric. Thank you for bringing this excellent to Devcon. What's the Obstacle of bringing the censorship resistant to the protocol level, what would be the Obstacle of, especially on the validator related.

B

Yeah, that's a that's a great question and there is a whole body of research that tries to, in an automated fashion, detect censorship and Implement like slashing attacks. Slashing responses in an automated fashion. I would say that the current to my understanding, I'm, not an expert in that either, but to my current understanding, there are partial Solutions. There are partial solutions that are still on the you know on the on the Whiteboard there's nothing.

B

We should probably expect that even even these partial Solutions will take a couple of years before we might see them integrated in the protocol level, so social slashing Remains the primary defense mechanism. For now it might be such that we have automated ways which would be much better right, because then it would be codified. Automated no subjectivity, no arbitrariness. So that's the goal right. So that's an open research area.

B

We might see advancements in that area that come underway, but currently we still have to rely on this social element uh where we are currently right now.

A

That's it. Yes, let's uh thank Eric.

A

Thank you seriously for your leadership on this issue. uh It's absolutely I think we might look back in Future years as this being the most important talk at this Devcon. um You know, I. Think the the all of the ethereum community is grateful that you that you're leading on us um or hopefully not- um and you know- because people get the idea and don't censor because they know they're going to get slashed, we'll see.