Tokenika's take on sustainable EOS arbitration

in #eos6 years ago

We have recently published our take on recent emergency freezing of accounts, and for the sake of clarity decided to do a more in-depth write-up on our vision of arbitration, so here it goes…

Tokenika logo 3d.png

Arbitration - the make it or break it in EOS

Since early on, we have been quite worried about the proposed arbitration bodies in EOS. We felt it was an extremely difficult task to get right, and it could be philosophically argued that in decentralised systems one cannot instate legally binding external rulings without disrupting the system’s decentralised nature.

That said, since there seem to be a lot of proponents for arbitration in EOS and some maiden verdicts seem to be out already, we would like to lay out a few guidelines which we think would help to make proper arbitration possible.

Source of the actual arbitration mandate

As a standby Block Producer, we are opposed to self-appointed, or instated through “friends or families”, type of arbitrators. We understand most of the community agrees with this viewpoint, but still want to stress how important this is. A make-shift solution could backfire on the EOS ecosystem, as it is not justifiable and brings a lot of bad press to EOS. Thus appointment of arbitrators needs to be a part of smart contract that is voted upon, as only then a given arbitrator has any mandate to operate.

Rules of engagement

In our view, 99% of the times ECAF arbitration should be willingly agreed to by both parties. The remaining 1% we will discuss later. Taking a step back, the word arbitration itself is defined as:

arbitration
noun [ U ] US /ˌɑr·bɪˈtreɪ·ʃən/

the formal process of having an outside person, chosen by both sides to a disagreement, end the disagreement

The current EOS ECAF arbitration alignment resembles a court in which we can unilaterally sue someone. In our view, this approach opens a can of worms, as the arbitrator engages on behalf of one side of the dispute, albeit without direct executive powers, held only by Block Producers. Currently, this might seem benign, as EOS does not process any real business volume, but at scale, when complexity hits, we see no applicable way to sustain the “blockchain police” model. Let’s imagine the intricacies of more complex cases. Suppose funds gets stolen and transferred further, taking part in legitimate transactions. Scenarios are numerous, and each one of them could turn the arbitration process into a legal spaghetti. Also note that freezing of accounts (executed by BP's) until resolution is found could be used as an attack vector, not to mention that any such intervention puts finality of EOS transactions in question.

Thus, back to the 1%, one sided “arbitration” resulting in freezing of accounts and other heavy interventions should be considered only if:

  • there is a legitimate risk for the integrity of the whole network.

  • 15 or more BPs agree that that is the case, and work together with the arbitration body, publish their intent, etc. The exact mode of operation could be ironed out, but the general idea of no overreach is the only thing that we see as workable.

Scale

Another thing worth mentioning is the volume of arbitrated transactions at hand — a cut-off point should exist, below which no case can be filed. This addresses a vector of attack on the network, where arbitration cases are purposefully too numerous to be handled by anyone. Think arbitration DDOS.

Recap

Only arbitrators elected through a smart contract by the community have any mandate to operate. This smart contract should be crafted carefully, and it is better that we take the time and effort to do this well, even at the expense of a fraud or two in the meantime.

EOS ECAF Arbitration should only apply if both sides agree to it. There should be no unilateral freezing of funds, except in most extreme cases, with network’s integrity at stake. There are many reasons for these constraints, from scalability to actual implementation of dispute-settling, when claims become too numerous and complicated.

There should be a cut-off point below which no case can be filed. This does not prevent people running their own multisig or escrow contracts and coming up with their own arbitration techniques, but that is their internal business and has nothing to do with EOS ECAF governance. We are sure there will be a big market for this kind of insurance and auditing agencies on EOS, where smart contract expertise and dispute resolution can present added value and create jobs in the system.

Below, see a short table, suggesting the mode of engagement of Block Producers and EOS core arbitration in problematic situations:

We, at Tokenika, see little room to manoeuvre around these constraints. As experienced crypto-libertarians and long-term DPOS advocates, we would be concerned if this was to be taken lightly, as it's the actual foundation of a stable building we're helping EOS to become.

Final words

This post is not meant to be inflammatory in any way, or question whatever happened so far. We believe everyone is acting out of their best intentions, will, and knowledge and the whole EOS community is performing a near-miracle, navigating this uncharted territory with a lot of grace. However, sorting these things out early on is very important in our eyes, as it orients the community in the right direction and sets up precedents for a sustainable implementation of EOS. We understand that some of our views may come across a bit radical, but think that the spirit of crypto was originally derived from self-responsibility and self-governance. We are open to discussion and will double our efforts to represent the EOS community to the best of our abilities.

Sort:  

Thus appointment of arbitrators needs to be a part of smart contract that is voted upon, as only then a given arbitrator has any mandate to operate.

I think this could be very problematic. If arbitrators are voted, then whales have huge powers in the system because they can make sure there will be arbitrators who they prefer.

No one should have more power than others to influence who are taking the case which they are part of. Who would trust a system where rich people can get judges for themselves?

Let’s imagine the intricacies of more complex cases. Suppose funds gets stolen and transferred further, taking part in legitimate transactions. Scenarios are numerous, and each one of them could turn the arbitration process into a legal spaghetti.

This argument has been made by many who are skeptical about the system. But they seem to think it will be blind and automatic, instead of intelligent people considering the case from all angles. That's why we have humans involved in the system, so we can get a reasonable outcome.

Also note that freezing of accounts (executed by BP's) until resolution is found could be used as an attack vector, not to mention that any such intervention puts finality of EOS transactions in question.

My guess is that account freezing will be quite rare. Scammers learn pretty fast that this is not a good place for them, and on the other hand, arbitrators require strong evidence for freezing. Of course they know there will be people who are trying to abuse the system.

One way to mitigate this is to require bonds from users who open a case. Bond should be so big that it can be used to pay for arbitrator's work and compensate other party if there has been false claim. It's highly unlikely that attackers want to use huge sums of money for this .

Another thing worth mentioning is the volume of arbitrated transactions at hand — a cut-off point should exist, below which no case can be filed. This addresses a vector of attack on the network, where arbitration cases are purposefully too numerous to be handled by anyone. Think arbitration DDOS.

Arbitration costs money. If there are a lot of cases, there is a lot of money to be made which will attract new arbitrators to do the work. DDOSing is also highly unlikely if arbitrators ask for setting a bond before they take the case.

As experienced crypto-libertarians

I count myself as crypto-libertarian, too, and that's why I like this system so much. This is the first blockchain that actually defines property rights and tries to enforce them. Creating a new jurisdiction that's not depending on governments is a huge step forward.

It must be remembered that ECAF won't be the only arbitration forum. It's just the first one. There will be free markets for all kinds of arbitrators and mediators which users can choose. And on the other hand, as Dan has emphasized, the priority is to make the system so safe and easy-to-use that people won't be needing arbitration. Arbitration is only for cases when everything else has failed, which should get rarer over time when the system evolves.

I think this could be very problematic. If arbitrators are voted, then whales have huge powers in the system because they can make sure there will be arbitrators who they prefer.
No one should have more power than others to influence who are taking the case which they are part of. Who would trust a system where rich people can get judges for themselves?

Actually what you're saying only proves the opposite point. As Dan Larimer has said, the role of arbitrators like ECAF is just to generate an opinion, not a final verdict, so that the ultimate judgment rests with BPs. And look what we have right now: Bitfinex, due to its impact on the top BPs, is the most powerful judge. Try to win a case against such a player.

That's why it's in the interest of all of us to have the power of the most influential players clearly limited by the Constitution. What we have right now is the opposite of that. We've granted enormous power to the richest. You might say they'll behave responsibly, as they value their reputation. But how is it different from PayPal?

The Constitution should be mostly about limiting the power of the most powerful. As it is now, it's mainly about everybody being subject to the unrestricted judgement of the most powerful.

This argument has been made by many who are skeptical about the system. But they seem to think it will be blind and automatic, instead of intelligent people considering the case from all angles. That's why we have humans involved in the system, so we can get a reasonable outcome.

Judgments in complex matters are... complex. What makes you think we can recreate on EOS a better justice system than the one we have in the real world?

I count myself as crypto-libertarian, too, and that's why I like this system so much. This is the first blockchain that actually defines property rights and tries to enforce them. Creating a new jurisdiction that's not depending on governments is a huge step forward.

Again, why do you think EOS obligatory arbitration has a chance to be a better solution than legacy systems?Just because we have "intelligent people considering the case from all angles"?

It must be remembered that ECAF won't be the only arbitration forum. It's just the first one. There will be free markets for all kinds of arbitrators and mediators which users can choose.

You're confusing arbitration to solve an actual dispute where both sides of the argument choose an arbitrator prior to entering a contract (in this case the competition between arbitrators is indeed a good solution) with "arbitration" to catch those who are accused of violating property rights (in other words, thieves and scammers). In the latter case, it doesn't really matter if ECAF is the only arbitration forum or not, as in this case ECAF acts as police and it's the victim who decides whom to employ, whereas the defendant has no choice in this matter. Imagine being accused of theft by someone connected to the top 21. You stand no chance of having a proper trial.

What makes you think we can recreate on EOS a better justice system than the one we have in the real world?

What makes you think we would be better off without a justice system? It's not like everything is always bad with existing justice systems and there is no upside to them. All systems are imperfect, especially where humans are involved, but some problems cannot be solved without human intervention. It doesn't mean we shouldn't try. I mean the argument that we should not do anything legacy systems do, just because we might fail to improve them, is wrong in my opinion.

All I'm not advocating writing off the justice & arbitration system completely. What I'd like to see is clear distinction between those two categories (as they serve different use cases) and applying the justice system only in very precisely defined cases. Rescuing people from theft is very tricky, especially when the ultimate judges (i.e. the top BPs) have disproportionate power and normal shareholders stand no chance against them.

Thus the primary thing the Constitution should do is define clear limits to the power of BPs instead of stating that all shareholders are effectively subject to justice served by them.

as the arbitrator engages on behalf of one side of the dispute

No - the arbitrator may be empanelled because of one only side, the arbitrator has to take both sides. Taking a snippet of RDR:

The Arbitrator is required to:

  1. Be independent;
  2. Maintain the balance of power between the parties,
  3. Give each party the appropriate opportunity to present their case and to respond to the presentations of other parties,
  4. Conduct a fair and efficient process, and avoid unnecessary delay and expense,
  5. Keep all communications in the case file.

These rules are typical of those found in all arbitrations around the world.

He has to take both sides, like every court which takes the defendants arguments into consideration, even when he's not there. This is not our point. Arbitration should be voluntary from the beginning and explicitly agreed upon by both sides of the deal. Not unleashed upon someone from atop, even with the best (stated above) practices. They maybe typical of all arbitrations, but these points are not the only ones typical. We argue that unilateral nature of the filed claim is in fact not arbitration but external ruling, resembling a court without executive power (since the arbitrator is not a BP, so no execution can take place).

You could argue that by agreeing to a constitution user agrees to ECAF. I think Dan has made this point somewhere and ECAF is definitely mentioned in a constitution. Both sides will always be EOS users, so both will have agreed to ECAF.

We argue that the construct itself is fundamentally flawed and will not scale. A blockchain police, imposed by the constitution or otherwise can only operate within an extremely limited scope if it is to be invoked unilaterally. Otherwise, we're importing the entire legal system onto the blockchain, and at this point, we see this as slightly delusional, due to the sheer complexity and sheer mass of the task.
We think it is much wiser to rewind the blockchain history of last 4-5 years, map out the most disturbing and loss generating events (heavy hacks, Ethereum DAO etc), and have the ECAF in place to address these. Not to try to serve everyone, including small investors that were phished, or have legal disputes/claims against others, rightfully or not. In our eyes this is already generating a massive dispute in the community, as the source of authority is questionable, the executive power is with the BP's (in DPOS by default), and everything becomes a bit of a castle in the sand, even if propelled by big ideas.

Scale

Another thing worth mentioning is the volume of arbitrated transactions at hand — a cut-off point should exist, below which no case can be filed. This addresses a vector of attack on the network, where arbitration cases are purposefully too numerous to be handled by anyone. Think arbitration DDOS.

This only works if the community is more or less equal. But it isn't. If we go to Africa and put EOS processes into the hands of informal workers they'll be facing cases of say $10. To them this will mean 5 days work.

Whereas if we look at the whales, they will happily DDOS the forum with cases above the threshold ... just to make sure it can't handle their $1mm case.

Life ain't so simple as the numbers!

We see a whale DDOS very unlikely, as large stakeholders have the most to lose. ECAF in our view is a special unit that should only interfere when network's health is at stake, or both sides of the dispute can agree and afford its service.

Please note that qualifying all cases of alleged theft as worth being considered, you effectively put the justice system entirely in the hands of private organisations like ECAF, as BPs won't be able to review them properly.

Chicken and egg - which came first!?

Recap

Only arbitrators elected through a smart contract by the community have any mandate to operate.

Like EOS mainnet, ECAF was started by interim bootstrapping. Which is to say, it cannot rely on the chain to elect its arbitrators because it was there before the chain, so the chain could start up with the BPs fully protected by dispute resolution.

Now as we move forward, nobody's comfortable ... but the reality is that we have no referenda contract so we're a bit stuck.

This point we agree upon, and have stated this in our previous post regarding the freeze. Still we think it's extremely important to avoid any further bootstrapped rulings, and focus on setting up a sustainable version. Make-shift state of this does more harm than good, especially from the standpoint of future investors and businesses looking into using EOS for serious operations.

very clear

Even a decentralized network requires a central recognition of principle or principles to be considered a reliable network