The State of Livepeer — January 2019

Some of the values that the Livepeer project emphasizes revolve around the subject of productive output: keep making progress, putting one foot in front of the other, iterating on a useful platform, and shipping. 2018 was a busy year along these lines, in terms of all that was accomplished in taking the platform from idea and early open development, to a live, functioning network. 2019 will be about taking this functioning network, and making it scalable and cost effective to meet the real world needs of scaled video developers. This post gives a quick look back at the highlights of the past year, and summarizes some of the focus areas we look forward to in the next.

Livepeer in 2018

Launched Livepeer to Mainnet
After running many internal and public testnets, completing audits on the code, and demonstrating the Livepeer software’s working use in video applications, the Livepeer protocol alpha went live on Ethereum’s mainnet on May 1, 2018. Activity on the staking and node operations side has been very high. Starting from 0% participation in staking, and very few token distributed, the network has quickly risen to over 26% of the outstanding LPT staked on the network. The 15 available transcoder slots have been competitively filled and operated reliably, and there have been a series of low severity issues identified and patched on the production network through iteration and a transparent community process. On the video side, although this has been an intentional alpha, the network has seen early demonstrated usage in non-scaled use cases like streaming events, meetups, and conferences. Livepeer has the ethos of launching early, making it real, and iterating collectively based upon user feedback and requirements. Projects will find, especially in the blockchain space, that when theory hits practice through a launch and live network, it’s far easier to iterate towards something of impactful utility than it is to attempt to design the perfect system pre-launch.

Invented, Launched, and Completed a New Form of PoS Token Distribution — The MerkleMine
In a stake based participatory network, like Livepeer, users need to have some token in order to participate in the first place. Livepeer’s goals were to set an initial genesis state for it’s network that allowed a very wide and open access to a large portion of crypto network participants (essentially all Ethereum holders), while also providing incentives and opportunity for those looking to actively run the network from the very early days to compete to generate a larger stake in the network while aiding in the distribution. Through a newly conceptualized mechanism, called the MerkleMine, Livepeer facilitated this over the course of five to six months, building up many hundreds of active participants in the network, and providing access to all who discover the project at a later date. While some have considered the MerkleMine to be one of the greatest form of PoS token distribution yet invented, and others have claimed that it is a total failure due to large gas usage and a plutocracy, the reality of course is that it is somewhere in between. It was an experiment that was open to all, and allowed anyone, on fair and equal terms, to aid in the distribution of LPT and generate a stake in the network to start participating. Livepeer hopes that other projects learn from this method, iterate on it, and consider the best ways to build communities and open access to their networks.

Conceptualized the Video Miner Proposal

One of the highest potential forms of disruption for the video infrastructure market is driven by the notion that there are idle video encoding/decoding chips on GPU cryptocurrency miners, that can be used to create additional revenue streams for miners without disrupting their current crypto mining. Livepeer’s coordinated marketplace can them to put this power to use, and the cost savings can drive down transcoding prices relative to the cloud by over 50x. See this video by Philipp Angele at Streaming Tech Sweden to understand the story as it relates to video, cloud, and mining in detail.

https://youtu.be/kpAv8VePX_8?t=1

Researched and Designed Streamflow Scaling Proposal
Livepeer’s alpha network work for simple streaming, and the providers running infrastructure on it may be charging very low rates, but the cost of using the blockchain is high, and the network isn’t yet reliable enough for scaled usage. Much of 2018 was spent on research and early development of the next release, called “Streamflow”, which addresses these issues with the alpha. It’s believed that the Streamflow design will be highly cost effective and scalable, and makes use of some new concepts like Probabilistic Micropayments, offchain job negotiation and failover, and open access to the network for far more than the alpha-capped limit of 15 transcoders. At the conclusion of 2018, a minimum viable version of Streamflow was being tested internally already.

Served as Case Study for Generalized Mining
An idea gained popularity in the second half of 2018, to an almost meme-like level of discourse, around the idea that if you actively participate in networks, you can accrue value and economics in that network beyond just the conventional buying and holding of token. Called “Generalized Mining” by some, “Mining 2.0” by others, and “Active Participation” by yet others, Livepeer was often cited as the example or case study of how this approach played out in practice in blockchain based networks. The reason that Livepeer was so predominantly fit the idea of this narrative:

  1. Livepeer was actually launched on mainnet.
  2. Inflation based participation incentives.

A whole cottage industry of staking as a service, validating as a service, staking based funds, etc had been growing in preparation for the launches of large public networks like PoS based Ethereum, Tezos, Polkadot, and Cosmos. But as Livepeer quietly launched ahead of them, with many of the same properties, it became a testing ground for how these ideas would manifest when the rubber hit the road. Analysis pieces like this one from Vision Hill Advisors go into far greater detail than this post can, but in short, unsurprisingly — if you actively add value to a network in alignment with the incentives and economic design, you will be better off than if you sit on the sidelines doing nothing. This is how it should work :)

Formed the Beginnings of a Growing and Engaged Developer Community
As a project who’s entire marketing budget over the first two years totaled only a couple thousand dollars aimed at supporting a few Ethereum ecosystem and Diversity and Inclusion focused events, Livepeer awareness amongst the general speculative crypto community remains lower than many ICO projects. However, with a strong focus on technical first execution, personal engagement with both the decentralized tech and video communities, and open and transparent communications, Livepeer has grown its developer chat to over 1000 users, gotten about 500 stars on its core Github repositories, had contributions from order 50 developers, and seen early development on top of its platform in a number of verticals. Applications like Scout have built on Livepeer, community miners emerged from projects like Bison Trails, community nodes and a nascent grants program have emerged in the US, in Hong Kong, and in Europe. And while the core team has put zero focus on any sort of trading or exchange while the network is in alpha, community members have put in efforts to ensure the token can be accessed in various platforms. Livepeer hosted the first decentralized video unconference this year, with attendance from many of the leading video platforms in the centralized world that power the majority of the world’s social and entertainment video content.

Livepeer in 2019

Scaling the Network — Launch Streamflow
On the technical side, the goal is simple: Ensure that the Livepeer network is scalable, reliable, and cost effective for video developers building applications at scale. This is what the Streamflow Proposal is all about, and the goal is to get it launched to mainnet as soon as it’s well tested, audited, and ready. The core team and community have been researching and building Streamflow for almost six months now, and a minimum viable version is being tested on a small private network already. This includes the first implementation of probabilistic micropayments, offchain job negotiation, and a scaling update that separates the role of Orchestrators from Transcoders. Following success of these internal tests, the team will continue implementing the full proposal, run public test networks, and prepare for a mainnet upgrade targeted around mid 2019. This should increase the number of actors who can play the role of Orchestrator/Transcoder from the current low order of magnitude during the alpha, to at least hundreds or more, further decentralizing the network and increasing its resiliency for end users.

Drive Adoption of the Livepeer Network at Scale
Early adoption of the Livepeer network has been meaningfully impactful relative to many other blockchain based projects, with groups using the network to broadcast from events like Devcon 4, ETH Buenos Aires, and numerous meetups and conferences around the world. Still, these are unscaled use cases often with only 1–5 streams at a time. As infrastructure, Livepeer really makes a difference not for one off live streaming, but for transcoding at scale — for applications that have hundreds or thousands of concurrent streams of video, such as large social apps (and hopefully DApps in the future), IoT connected devices like cloud-enabled security monitors, and industrial video use cases. As Streamflow launches and the network is reliable and cost effective enough to service these use cases, the project will shift energy purely from development, to also driving adoption and bringing users online through pilots, trials, and conversion of their infrastructure to the public network on mainnet. While much of the attention around Livepeer has been on the blockchain/crypto side during the bootstrapping phase, it is really in the video world where Livepeer delivers impact and provides utility. Expect this side of the project to shine bright in 2019. As the network achieves delivering over 10x cost savings versus cloud services (initially, and up to 50x over time), in what is an exorbitantly expensive market, the belief is that Livepeer can be the first token coordinated, decentralized resource allocation protocol to deliver scaled value and impact.

What Else?
There’s a long list of areas that Livepeer will continue to improve upon in the coming year, including growing the developer community, pushing the bounds of decentralized governance and development through community nodes and inflation funding, and continuing research on additional areas that can leverage decentralization to bring disruption to video infrastructure such as content delivery, verification, and stream processing. But as mentioned above, there are really two areas of primary focus that everyone in the community should prioritize and rally around:

  1. Scale the network
  2. Drive adoption

If we all keep these two things in our sights, then 2019 will be an exciting, challenging, and rewarding year for those choosing to spend their time collaborating around the Livepeer vision.