Livepeer Decentralization Update

The Livepeer ecosystem has grown tremendously since launching to main net one year ago. The number of independent participants who operate the network, develop the core software, and build applications on top of the Livepeer continue to expand and diversify. This post showcases many of the entities building and operating Livepeer, as well as highlights the path to even greater decentralization.

Decentralization of Network Operations

Amongst the various vectors of decentralization, the operations of the Livepeer network itself is likely furthest along the spectrum towards the decentralized end. There are over 1600 accounts participating in the network by staking their LPT, contributing to an over 40% participation rate. Amongst the 30 transcoders elected in the protocol explorer, there is a cross-section of geographies, expertise, and differentiated value propositions. The company that employs the core team, Livepeer Inc., doesn’t actively run any nodes itself, and core team members are independently involved in running only 3 of the 30 transcoders — often collaborating with independent parties in the process.

The underlying economic protocol is secured by the Ethereum blockchain, which prioritizes decentralization and security at the expense of performance — and makes it such that any economic attack on the consensus-based data in Livepeer would require an attack on Ethereum itself — a very expensive endeavor.

Many of the operators of Livepeer itself bring significant experience across both the blockchain and video industries.

Companies like Bison Trails, Staked, and Figment Networks offer professional level infrastructure as a service across multiple blockchains, and are active participants in the Livepeer network on behalf of themselves and their own business relationships. More independent projects like Chainflow also operate across networks, and encourage organic community-based staking models emphasizing decentralization, and often achieve performance and reliability at par with some of the bigger, better funded infrastructure operators.

Video encoding companies, like Qencode, operate nodes on the network in order to utilize their deep expertise in video infrastructure and position themselves to win future work on the network. These forward thinking companies look at decentralized markets as an opportunity to expand their own addressable business beyond closed, centralized transcoding infrastructure.

Other nodes have focused on geography-specific coverage, such as the Asia/Hong Kong community node, a Finland based node focused on northern Europe, and nodes advertising coverage for high demand regions like US West. Geography should expand significantly with the Streamflow release as well, expanding to 100+ active nodes, with fee and performance based incentives for global nodes to serve local demand baked into the protocol.

And finally, there are many operators on the network who are also building community and video software projects around Livepeer — investing their time in the model whereby their contributions to the ecosystem actually increase the utility of the underlying network in which they are stakeholders and operators. More on these in the next section.

Decentralization of Development

The Livepeer media server, node, and protocol are all open source and fully available on Github. While in the early days, Livepeer Inc drives much of the video specific technology going into the transcoding platform, 3rd party development around the ecosystem has started to grow significantly, with over 50 external participants in the open source core development. In addition, here are some of the highlights of 3rd party development around the ecosystem.

Epic Labs, an innovation lab out of Spain that also contributes to the Swarm protocol, drives much of the research and development around video quality detection and non-deterministic transcoding verification. They develop the Lightflow video platform for AI optimized video streaming, and have been great supporters of decentralized technologies like Livepeer.

Scout.cool develops the primary network explorer that network participants use every day to monitor transactions and view protocol statistics, and partially fund this effort through the community inflation funding routed through the protocol via their node’s reward cut.

The primary staking analytics and alerting applications are also developed by independent parties such as Livepeer.studio, Livepeer.tools, and the Telegram Bot campaign from community member vires-in-numeris.

The explorer staking tool, which serves as the gateway for staking and delegation, is undergoing a Streamflow facelift from a Community node organized team. And the team from Consensys company, 3box.io, is integrating identity features into the explorer to aid users in assessing and electing transcoders.

While much of the ecosystem has been developing tools on the infrastructure and supply side of Livepeer, Livepeer Inc has focused more time on building product and services on top of Livepeer necessary to drive adoption to the protocol amongst video platforms and applications.

One of these applications on top of Livepeer, DTok, a decentralized pay-as-you-go version of Tik Tok, is being built by the community and provides end users with the ability to broadcast video, have direct economic relationships with their users, and leverage Livepeer’s underlying video infrastructure.

And finally, some experimental decentralization oriented developers are running nodes and campaigning around governance projects — things like the VideoDAC and Commonwealth, that aim to let decentralized groups of participants operate infrastructure collectively.

Decentralization of Community and Governance

As observed in some of the challenges that other blockchain ecosystems have faced around governance, decentralized coordination is a hard and unsolved problem. For now, the community always has the opportunity to fork the protocol in the case of a governance dispute. But as noted in Livepeer’s Path To Decentralization, there is a roadmap to move towards decentralized governance of parameter and protocol updates. While this isn’t yet encoded in the protocol itself, there have been positive steps taken in the first year of operations.

There have been a number of protocol updates deployed which went through a transparent community input and review process. The Streamflow protocol update sought input from over twenty decentralized tech ecosystem external reviewers, and it received opt in support from over 85% of the actively operating infrastructure providers on the network.

There is a community committee that self forms to provide bi-weekly community calls, grant moderation, and user support+moderation of the Discord chat room. Community members are developing primers and micro-sites to educate new users, and are driving much of the feedback and user research process that brings information into the project as a whole to inform the path forward.

One of the things that the community has its eye on closely are the decentralized governance mechanisms that are beginning to play out live in ecosystems like Cosmos, Tezos, and Decred — routing community funding, and enabling protocol upgrades without any centralized decision making at play. In a post-Streamflow world, as there are over 100+ orchestrators on the network, and parameters may need to be adjusted over time to optimize the networks performance, security, and trust, it will be necessary that these actions can be taken by the community stakeholders themselves, without risking a divisive fork.

Challenges and Conclusion

Decentralization is a tool that the Livepeer project uses in order to deliver tremendous practical benefits to the video and web3 ecosystems. But decentralization isn’t easy and doesn’t come for free. Some of the challenges that have to be overcome along the way include:

  • Tendency for stake to centralize and ossify in the delegation protocol.
  • Growing a community of maintainers and active contributors to the Livepeer Media Server core video software.
  • Balancing the protocol’s ability to quickly patch critical bugs with slower moving decentralized parameter and protocol updates.

The community is aware of the above challenges, discusses them frequently, and is working together positively towards proposals, experiments, and solutions to tackle them head-on. As has been observed across many blockchain ecosystems, it is easier to coordinate around solutions to the tough problems when the community is just beginning to grow, than it is later on after the project means many different things to many different stakeholders. The Livepeer project’s transparent, open, development focused, culture has allowed alignment on many of the tough issues so far, even as the project has decentralized significantly — but there are still a number of challenges ahead, and hopefully by being aware of them going in, we can overcome them together along the way.

Note: If you are interested in participating in the Livepeer project and joining the community, then you are welcome to join one of the upcoming bi-weekly community calls. Discussions around decentralization occur frequently.