Is DPOS the Next Step in Distributed Consensus?
A Short, Simple, Logical Perspective on EOS
I have found myself discussing the future of consensus on a number of occasions, with a diverse group, and the sentiment that many seem to share is that delegated proof of stake (DPOS) is the next step in the evolution of blockchain, or the 3rd wave of blockchain-tech as some are saying. I would like to dive in and open up my current thinking on the subject and challenge my own ideas, and leave others with the chance to formulate their own conclusions if this is the way forward, or simply a short-term creative competitive advantage.— Kyle Forkey
Breaking It Down
Let me first give a brief overview of DPOS. This is a consensus method by which coin holders have the ability to vote for delegates to run nodes which secure that coin’s particular chain. These delegates are then expected to — honestly and fairly — validate blocks, lest they find themselves voted out of a job by their fellow validating constituents. The key advertised benefit to potential developers being that these nodes will run enterprise level hardware, while only having a handful of nodes to maintain the ledger with accordingly. This enables transaction speeds not currently possible on traditional blockchains, but sacrifices *some* of the core tenants of decentralization. But, is this progress, or a creative means to bootstrap an alternative smart contracting platform; one capable of challenging the growing behemoth that is the Ethereum developer community? Let’s dig a bit more.
The Importance of Sustainable Decentralization
There is an inherent centralized nature with regards to delegated proof of stake, when compared to how traditional blockchains have been secured, historically. We are sacrificing validator node diversity and a slow crawl towards further real decentralization, for a system in favor of transactional speeds and making an incremental step towards evolving the concept of mining; which to myself, seems like taking a sizable step backwards.
Why then, would 4th wave tech not just be the removal of multiple nodes entirely for the fastest consensus possible? There is a project currently processing north of 60,000 transactions/second through a single node, and its owned by Visa. It seems to me that the fewer nodes we need to maintain a ledger the faster these systems can operate; though the final product is at best redundant, and at worst useless. I would like to explore the outcome of a protocol structured in this way.
The first order of business in a delegated proof of stake system is to find your validators, the individuals that will secure your blockchain and notarize your blocks accordingly. Using EOS as a popular example with its planned 21 block producers (delegates); if a group is going to invest in the infrastructure to secure the EOS blockchain, it would seem foolish to leave it to pure chance that they are still producing a year later. That group is certainly going to be willing to cut into its margin for the assurance that they remain an active delegate. Think any politician whose ever been up for re-election, and the measures they take to stay in office. How then, will delegates go about securing these votes? In steps the vote cartels who supply the required votes in exchange for a share in the profit. These cartels will operate similar to the crowdfunded, cloud mining farms of today — by way of sending money or the coin required, you will be able to participate in profit sharing with that cartel who in turn gains even more negotiating power, possibly enough to pick their own nodes.
It follows that these vote providers would amass a majority of the votes as it makes little sense as a voter not to participate and receive your due payment. Even if the majority was repulsed by the idea of accepting payments, the minority would have greater economic incentive to press forward than they would to press back (an EOS that pays is worth more than an EOS that doesn’t, so I will buy yours at a premium).
Interestingly, we seem to have arrived back at a proof of stake system, less the ability for individuals to directly participate in consensus. Vitalik wrote a piece on this you may find interesting which dives into the driving motivators in a chain such as this once voters are being paid in regards to EOS’s seemingly emerging plutocratic governance model. A surprising rebuttal to this chain of thought is that this behavior is simply not allowed; which seems akin to saying politicians aren’t bribed because it isn’t allowed. I do not think this amounts to even a bump in the road for bad actors.
This tweet by Nick Szabo sums things up nicely:
The next logical step in this system is that an individual, group or groups with aligned incentives, would seek to gain full control of the chain through a simple majority (you can learn more about collusion and oligopolistic game-theory from this very insightful article by Policonomics, here). This can be accomplished through normal free market competition. If one delegate finds a way to increase their margin (proprietary ASICs, cheaper electric, etc…) they can offer higher premiums for votes thus increasing the number of nodes they control. This does not mean however that this group would steal, double spend or do anything else to harm the chain as that would destroy their investment. It does mean that they decide the rules of participation on this chain. This includes royalties, fees, penalties, terms of service and whatever else they deem necessary. It also means that they get to decide what the rules are in regards to voting.
At some point, the cartels themselves would be in danger of extinction due to the fact that the behavior of paying for votes would become public and expected and probably framed as “shared ownership of the network.” Of course the cartels would then with-hold votes for delegates, attempting to cut them out and ultimately merge with them in a sort of partnership. This would be a tumultuous time, or maybe it was this way from the start; but your final product seems to be the cartels and miners becoming one, and paying users directly.
The most interesting development in this entire process is what comes next: Having a unified body with complete control over how this chain operates, they now influence the rules by which that chain operates, and are beholden to no-one but pure profit making. They now have no incentive to pay for votes, and will instead focus on profiting from users (because they are a business). Then follows the inevitable knee jerk “Hard fork now! We must resist!” (minorities can be very powerful forces) at this conclusion; but chains follow the miners, and developers will be become ensnared in their social walled garden and assurance of “app scalability”.
If the entire chain is secured by one group there will be no successful forks barring mass dissent within the delegates themselves. Not to mention this chain will be fast in comparison to any potential forked chains who try to start this process all over, and isn’t speed the main draw to this system in the first place?
This system now bears a stark resemblance to PoW minus the W. You now just have a corporation running a centralized “blockchain” that you can build “dapps” on. Why not try AWS? It’s pretty great if you haven’t heard of it.
Decentralization Going Forward
You may disagree with my conclusions, but the reason this system falls apart is because you have created an system in which incentives can be paid for. The cost in becoming a delegate is high, delegates will do what they need to mitigate the risk of loss through buying votes. Some users will undoubtedly be willing to sell their votes. There are individuals whose primary source of income may even lie in voting, which would be a valid economic contribution except for the fact that they aren’t the ones voting. What voters will be doing by accepting payments for votes is giving up their power to stop bad actors from doing bad things.
It stands to reason a similar scenario plays out in any system whose cornerstone of advancement lies in concessions to decentralization. Individuals will always act in their own self-interest. The only hedge to this being your own self-interest. If we don’t have a say in the system we participate in, what have we gained? If as an industry we aren’t quite as decentralized as we advertised, yet, perhaps we ought not lose sight of the ethos that got us this far in the first place.
One of the core tenants of any crypto investment thesis and focus should be built around sustainability. Have we reached the point already, that we’re willing to centralize wealth creation and developer ecosystems into the hands of another dominant minority? Will we learn a different lesson this time? Or, is this just a temporary step in our journey towards the trust-less, fair, sound, and decentralized future that our collective original ethos was built upon? We look forward to finding out, together.
For other varied perspectives (always do your own research to formulate your own mental models and opinions), check out these posts.
Vitalik Buterin’s Thoughts:
Governance, Part 2: Plutocracy Is Still Bad
Coin holder voting, both for governance of technical features, and for more extensive use cases like deciding who runs…
Notes on Blockchain Governance
In which I argue that "tightly coupled" on-chain voting is overrated, the status quo of "informal governance" as…
Other Various Opinions:
The Limits of Crypto-economic Governance
Vitalik recently made claims that Delegated Proof of Stake (DPOS) results in rule by plutocracy (government by the…
EOS Is Coming, If Anyone Can Figure Out How to Vote - CoinDesk
After a year-long initial coin offering (ICO), what has perhaps become the most hotly anticipated blockchain is…
EOS Analysis and Valuation - Multicoin Capital
We recorded a podcast and webinar discussing our EOS report and answering questions from the community. To listen and…
Edited by: Steven McKie