In the quest to find a superior consensus mechanism, Delegated Proof of Stake (DPoS) seems like a winning bet. In brief, DPoS stakeholders elect what are called witnesses to generate transaction blocks and secure their blockchain network. Stakeholder voting strength is proportionate to the number of tokens any one stakeholder owns (thus, those with the most stake have the most say).
While DPoS rewards the top 100 witnesses with tokens, the capacity to remain a witness declines as competition for these slots intensifies (stakeholders elect witnesses based on current # of block producers, geographic decentralization, and even the capacity for witnesses to remain policy neutral). As such, witnesses can be considered employees of the stakeholders.
Why the emphasis on witnesses? A dedicated and limited number of trusted block producers can process transaction blocks much more efficiently. By creating consistent and reliable block production, witnesses enable DPoS networks to become much more scalable. While the human element plays well here, it appears to be somewhat inferior to infinitely more scalable propositions (such as sharding).
Since DPoS uses trusted witnesses, it’s far less vulnerable to the “nothing at stake” issue inherent in Proof of Stake schemes (where PoS validators faced with a fork and having ‘nothing at stake’ find it profitable to validate both chains). However, DPoS coins remain at risk from block producers paying for votes or colluding for nefarious purposes (i.e., double spends or censorship). Whales (or a group of whales) may also seek to take over the voting process (though DPoS voting incentives give them undue influence as it is). Stakeholder apathy encourages such attacks and remains a significant weakness in DPoS.
DPoS founder Daniel Larimer asserts that both Proof of Work and Proof of Stake are prone to further centralization (as the former pushes out smaller, less technologically proficient miners and the latter pushes out those with smaller stakes via higher transactional costs). As he tells it, DPoS was chiefly founded to correct these problems.
That’s not to say DPoS isn’t prone to centralization issues. However, DPoS proponents assert that small compromises on decentralization help to ensure better scalability and security. The founders of DPoS thought it a fair tradeoff, as evident in how they introduce it as such on GitHub,
Unlike other methods of securing cryptocurrency networks, every client in a DPOS system has the ability to decide who is trusted rather than trust concentrating in the hands of those with the most resources. DPOS allows the network to reap some of the major advantages of centralization, while still maintaining some calculated measure of decentralization.
At first glance, DPoS appears to be an elegant solution to the technological tradeoffs inherent in current consensus methods. However, their governance model seems to short-circuit voting incentives for all but whales. As even a pro-DPoS whitepaper explains:
[Exploiting low voter turnout] is the most obvious attack against a DPoS blockchain. The core of this attack is the fact that in any voting system, very few participants actually show up and vote. In blockchain token voting systems, anyone with a small stake is unlikely to actually influence the direction of the platform with their vote. The time spent researching on how to vote may not be worth the effort for what they view as a minimal impact. Still, the natural result of these systems is often that overall voter turnout is low, and voting is mostly done by whales, exchanges, and wallet providers.
Given how DPoS voting is structured, whales play a naturally pernicious role by default. In Lisk, for example, DPoS consolidates the power of a few select delegates to create an oligopoly (as revealed by several graphs in this article). Its DPoS scheme exacerbates “the rich get richer” tendency already inherent in current financial systems. Not incidentally, Ethereum founder Vitalik Buterin asserts as much in a short but brutal Reddit post written less than a year ago.
Unlike DPoS, XTRABYTES’ Proof of Signature (PoSign) consensus mechanism does not rely on mining. Rather than witnesses, PoSign relies on a fully decentralized network of STATIC nodes to instantaneously verify transactions (employing a proprietary alert system named PULSE to communicate with network nodes as to when a transaction block should be verified).
Moreover, a PoSign consensus mechanism avoids the security vulnerabilities that witnesses or apathetic stakeholders present. PoSign’s STATIC nodes (all 512 of them) sign off on every transaction and can instantaneously recognize a security threat. Expect PoSign to become a formidable challenger to DPoS soon.