What can Livepeer learn from Cosmos? Considerations for Decentralized Governance
On September 12, 2019, Chris, Lorien, and I discussed governance with the Livepeer team and community. Learn more about Livepeer staking here.
Roadmap for Decentralization
After decentralizing ownership of the network by way of the LPT token (via merkle mining and inflation), Livepeer Inc. is preparing to release Streamflow. Streamflow is the scalability and ‘end user adoption’ focused upgrade to the Livepeer network, and is expected to enable Livepeer’s scalability on Ethereum with three components: orchestration, probabilistic micropayments, and off-chain job negotiation.
What’s the next step for Livepeer after Streamflow?
Decentralized governance is next on Livepeer’s decentralization roadmap. There are three phases in the roadmap:
- Decentralize the stake (LPT) in the network (merkle mining, inflation)
- Decentralization of governance (comes after Streamflow)
- Full decentralization of the tech itself eg. storage to guarantee availability of files
Identifying Livepeer’s Governance Needs
Nelson Ryan identified a potential need for branches of governance:
1) code & upgrades; 2) funds and resource allocation; and 3) decision-making control.
Beyond social consensus, Doug Petkanics noted a need for effective execution of what’s been decided upon, including checks & balances.
How incentives are set up and controlled, via governance, can help to ensure the ongoing execution, outcomes, and the care that goes into these networks.
What will Livepeer strive to decentralize and how can governance help?
Breaking It Down
Livepeer will be looking for a community consensus based way to set parameters that affect the economics of the network. For example, the Livepeer community should have the ability to change the size of the active transcoder set. Governance mechanisms may be used to find, implement, and execute this sort of consensus.
Code updates to the network should be derived from community consensus, and again, governance mechanisms may be used to find, implement, and execute that consensus.
Network defences, such as protocol pauses and bug fixes, will need to be executed quickly in order to safeguard funds, but without introducing single points of failure. The processes for these defences should arise from community consensus, and perhaps decentralized governance structures could be used to achieve these goals.
According to Doug, these components will be critical to Livepeer’s decentralized governance phase, once the Livepeer network can demonstrate its utility.
Models of Governance
Something that I found interesting was a question about establishing a network constitution ie. a core value set. Is it too late for Livepeer to refer back to a core set of values for a governance framework? My thinking: be skeptical of establishing a core set of values beyond technocratic ones, because our values tend to change as the ownership of the network evolves eg. with new stakeholders.
It’s difficult to find an agreed-upon interpretation of a situation in relation to the constitution, in that the conversation tends to become a debate about whether one point of view better reflects reality than another point of view.
Defaults weigh heavily, so it may be difficult to revise a constitution to better serve a changing network. However, a guiding framework could bring less contention and be more flexible to evolve as new stakeholders join the network.
Doug was also curious to know if social consensus ever involved referring back to the original Cosmos whitepaper.
Observing Other Networks
Chris and I testified that, for the most part, off-chain governance discussions among Cosmos validators rarely seemed to refer back to the original whitepaper. The only example I could think of which may have applied, was the decision to increase the active validator set–though the discussion did not seem to centre around plans from the original whitepaper. For the most part, Cosmos Hub governance decisions are determined by the voting power of its validators.
The pros/cons of bundling network security with network control
Since token-holders tend to be passive in token-based voting, Cosmos validators run the network and by default control the decisions to change to the network. That concentrates a lot of power in their hands. Question the role of validators or transcoders in governance. Perhaps there’s an opportunity to better distribute that power.
Validators tend to be vigilant, which is convenient, because attention is necessary for good governance and token-holders seem to be passive. But that doesn’t mean that validators’ values are ideally aligned with the values of the network and the other stakeholders. Each may pay attention to different things.
How does bundling governance with security change incentives?
It could be more valuable to a validator to have governance voting power than income for providing security. This may dilute or drive down the value of security. An example of this is the Cosmos validator Sikka, led by Sunny Aggarwal, who appears to be leveraging a zero-fee commission and his insider status to dominate the validator set’s governance voting power.
However, transcoders are different than validators, in that they provide security and a transcoding service to the network. Their values may be better aligned with those that use Livepeer’s network than those who use Cosmos’ network.
I think that there should be a plan to build capacity for a diverse and growing set of stakeholders to participate in governance before establishing an elaborate governance system. I strongly agree with Doug, in that I think that the network should be functional and useful (ie. worthy of governance) before transitioning control into the hands of the community, and so I like using technocracy to drive initial development.
If this has piqued your interest, I suggest listening to this audio recording of the community call that inspired this article:
Hopefully you found this useful. Feedback is always welcome! I’m on Twitter.