Ethereum 2.0 is Prime CryptoTulip Material

Ethereum 2.0 will be a strong candidate for CryptoTulip of the Year (though maybe not this year)

Ethereum was the winner of the first Cypto Tulip of the Year Award, and a perennial favorite. This year may see Ethereum begin to move beyond Nakamotoan consensus, to a Proof of Stake protocol. This would will the biggest PoS system ever tried.  It is far from certain when, if ever, Ethereum 2.0 will go live. But if this goes through it will be radically non-Nakamotoan, and I think it would mark the beginning of post-Nakamoto age!

What’s So Great About Proof of Stake?

There are several radical innovations in Ethereum 2.0, but the biggest is the use of Proof of Stake verification, which replaces Nakamotoan Proof of Work (mining).

Vitalik Buterin, the charismatic leader of Ethereum, has been interested in Proof of Stake for a long time now, for very good reasons [1].  Classic Nakamotoan Proof of Work is heinously wasteful—it’s very logical principle depends on being wasteful—and has led to consumption of vast amounts of electricity, often from the cheapest, dirtiest sources.  Proof of Stake (PoS) replaces this brute force nonsense with a much less costly process that is supposed to do the same thing.

I admit that I don’t fully grok PoS, but the basic idea is that a subset of the network is picked to be the validators, rather than having the whole network do it.  To be a validator, a node puts down a stake, which it will lose if misbehavior is detected.  The selection is influenced by voting, which is influenced by reputation.  So, instead of open competition based on the amount of computing power you can amass, there is more organized competition based on having money to bet.  Scoffers note that this is a form of plutocracy, far from Nakamotoan peer-to-peer equity.

The other notable point is that the network trusts the selected validators.  Obviously, everyone can check the work, but the validators are the law, unless fired (which is pretty drastic).  Furthermore, the small number of validators is, in a key sense, a centralized topology (which is why you have to trust them).  And, by the way, their word is final, which is a somewhat subjective definition of the “immutable” state of the system.  All this is not Nakamotoan.

There is some complicated economic modelling underpinning this approach. I haven’t seen any simple explanation of what these models are or how they have been validated.  Buterin seems to have at least some understanding that abstract game-theoretic models are not necessarily complete or even valid in the real world.  So one wonders why there is a need for so darn much complex theorizing.

(Personally, I would strongly recommend publishing some detailed technical papers in a peer reviewed  process, so debug all this stuff.  But what do I know?)

Philosophy aside, Proof of Stake in general and in specific implementations has not been widely tested or critiqued.  It is completely different from existing Proof of Work systems, so previous experience is not much help.

What Could Possibly Go Wrong?

This winter, Pascal Thellman writes about a number of potential weaknesses in P0S “validators” [2].  It is important to note that no amount of theorizing will tell us how well a particular implementation and deployment stands up to these kinds of challenges.  We need to really test the real thing, which is not a trivial undertaking.

Theilman identifies two major areas of concern:  ‘attack vectors’ and ‘misaligned incentives’.  Such ugly, technical topics. Already, I hate it.  Let’s see what he is talking about.

The ‘attack vectors’ are mind bending, and cool in the way of all great hacks.

Theilman suggests that, in principle, it might be possible to obtain private keys that were used in the past to validate records, and use these to rewrite the history on the blockchain.  Cool!   He says this shows that the PoS blockchain is only “temporarily” valid.  I would say that it shows that securing the validity of the blockchain requires securing every validator that ever contributed to the records.  Good luck!

He identifies a “Sour Milk Attack”, in which malicious nodes fill the network with deceptive bad records, obscuring which nodes are well behaved.  Anyone who thinks this kind of information warfare isn’t a real threat should have a look at headlines about election meddling.

He also says there are other attacks that can mess up, crash, or overwhelm legitimate validators. I’s say that the “centralized” design must, by definition, open the network to denial of service attacks on the validator nodes.

Setting aside the fact that I am not the only one who doesn’t understand the fancy incentive systems, the basic design is obviously open to many simple misaligned incentives.  Wealthy players will be able to grab the rewards, leading to concentration and control in the hands of a few.  Furthermore, there seem to be various possible loop holes and wrinkles that can be played by validators that might mean they can game the system to get profits improperly or unfairly.

He comments that much of this complex incentive stuff “highlight[s] the layers of abstraction needed to obfuscate the fundamental issue of staking without actually addressing the problem itself.”  Quite!

I’ll add to Thielman’s points.

To date, many PoS implementations have be pretty opaque about how they work.  Classic Nakamotoan consensus works in part because it is so brainlessly stupid and therefore simple to understand.  There are obvious problems (e.g., concentration of mining power), but they are very obvious, and not obscured by fancy game theory.

Is this a Future Crypto Tulip?

With Buterin’s determined support, it seems likely that Ethereum will push forward on development of Ethereum 2, with PoS inside.  Simply getting it up and running will be difficult, and getting everyone to accept the new version may be harder.  The good news is that Buterin is working slowly and carefully, so there will be a lot of interesting testing before it goes live.

I think Ethereum 2.0 could be the Crypto Tulip to end all Crypto Tulips!  If we can swap out Proof of Work mining (and add sharding and other novelties), changing the very meaning of “immutable”, “decentralized”, and “trustless”, and still call it blockchain technology, that has to be a strong candidate for CryptoTulip.  Tulip 2.0!

Stay tuned!

---

1. Vitalik Buterin, A Proof of Stake Design Philosophy, in Medium. 2016. https://medium.com/@VitalikButerin/a-proof-of-stake-design-philosophy-506585978d51
2. Pascal Thellman (2019) Validators Create New Attack Vectors for Decentralized Systems. Coindesk, https://www.coindesk.com/validators-create-new-attack-vectors-for-decentralized-systems

 

Cryptocurrency Thursday