Livepeer released the latest version of its software and protocol today, codenamed Tributary. This was the first major release since the protocol launched to Ethereum’s mainnet on May 1st. The goals of the release were to enable video developers looking to use live video infrastructure services at scale to reliably begin to build on top of Livepeer’s video transcoding network, and to address the largest pain points that protocol actors and token holders felt in the initial alpha. The release contains the following features:
- Partial Unbonding — Token holders who bond their tokens can now unbond some of their tokens, without having to unbond all of their tokens. This will let the nodes operating as DAO’s, development projects, and community projects begin to use some of the token they’ve earned for development and growth.
- Re-usable transcoding jobs — If you would like to transcode video using the same settings as previous jobs, just re-use the previous job, and avoid having to pay additional ETH fees to put the new job on chain.
- Reliable direct connections between broadcasters and transcoders — this key feature will dramatically increase reliability for developers beginning to test Livepeer’s transcode network.
In addition to these high level features, there are a number of protocol improvements which were run through our public LIP process, described here, as well as a number of Livepeer node updates described here.
If you are running as a transcoder, or a developer building a video enabled application on top of Livepeer, see how this update affects you here.
The Tributary release was intended as a “fast-follow” after the initial protocol launch, in order to address major pain points and add highly requested features. But going forward it’s clear that reliability, scalability, and cost effectiveness are the key features which will take Livepeer from an alpha, to a suite of usable live video infrastructure services for developers and scaled media applications. As such, we are already hard at work on the research, development, and testing of designs that address the following:
- Enable scalable, cost effective GPU transcoding, via the Video Miner proposal. Create a two tiered system allowing current transcoders to “orchestrate” jobs that are competed for via less protocol aware GPU miners and spare hardware.
- Expand the network and create competition for available work — currently jobs on the network are assigned to transcoders based upon stake weight, regardless of their location or capabilities. Create mechanisms that grow the network footprint and capabilities, allowing developers to access the services they need in the region their video originates.
- Introduce additional live video infrastructure services leveraging decentralization that developers need including p2p content delivery and storage options. We’ll integrate with additional layers of the web3 stack that help to fulfill these needs for developers, but focus on customizing for a live video context.
- Continue down Livepeer’s Path to Decentralization, on the fronts of infrastructure, governance, and network ownership.
While Tributary is a significant milestone that should make developing on Livepeer’s alpha much more practical, look for a v2 architecture proposal published in the coming months, which addresses many of the above to make Livepeer a scalable and cost effective option for developers.
Thanks to all the contributors to the Tributary release — the core team, the open source video community, and all participants in the Livepeer network who provided feedback and requests. Onwards!