The Streamflow Proposal — Scaling Livepeer
The Livepeer core team and community have been hard at work on a proposal called Streamflow to upgrade the current Livepeer alpha to be more reliable, cost effective, scalable and performant. Today, we’re excited to share this proposal for community feedback and review. These updates will make it possible for video developers to shift their infrastructure to the Livepeer network and realize cost savings. This post provides a summary of the paper, and how the updates impact you as a token holder, transcoder or developer. If you want more details on the proposal, read the Streamflow Paper here.
Streamflow introduces some new and changing concepts to Livepeer. Some of the more noticeable ones to all users will be:
- Offchain job negotiation rather than on chain job assignment — major improvements for video developers using Livepeer, as they can now work with multiple nodes for redundancy or switch quickly in the case that the transcoder they are working with becomes unavailable.
- Payments via probabilistic micropayments (PMs) — this means that nodes won’t need to do 5 transactions on chain to get paid for each job, but instead can just cash in their earnings whenever it is affordable to do so on the Ethereum blockchain. This is a major cost reducer, and can make the Livepeer network very cheap to use for developers regardless of the fluctuations in price of Ethereum gas.
- Relaxation on the limit of number of active transcoders. Currently this is arbitrarily 15 during the alpha, but in Streamflow, the number of nodes who will be able to provide enough security in the form of stake to compete for work on the network will increase by an order of magnitude.
- Introduction of a new role called Orchestrator, which coordinates work amongst a pool of transcoders for further reliability and better devOps setups, enabling GPU transcoding pools.
- First-level verification responsibilities shifting to the Broadcaster, avoiding the need (and cost) to invoke Truebit unless a fault is detected, and avoiding the data availability problem.
While early usage has been focused largely on demonstrations, streaming of events and conferences, and other single-track usage — these updates will make Livepeer viable for video developers with scaled applications, with 50+ concurrent streams, to shift their infrastructure to the network to realize the cost savings that can be unlocked.
What does Streamflow mean for different types of users in the Livepeer ecosystem?
Many of these updates are Livepeer specific concepts, but the Probabilistic Micropayments mechanism is a general scaling technique that works for many tit-for-tat use cases like streaming, and we’re excited to develop open source libraries and specifications that let other Ethereum based projects benefit from the tremendous costs savings this can enable.
Token Holders Staking Their LPT
Token holders have the responsibility of staking towards effective nodes in order to QA them and route work their way. Orchestrators still offer reward cut and fee shares. But whereas much of the incentives in the alpha network have focused around the rewards, in Streamflow fees should start flowing more prominently through the platform in ETH. As there will now be far more than 15 nodes, you’ll have a more active role in optimizing your own fee returns across a number of competing candidates performing a higher volume of transcoding work — essentially more proactive review and delegation will be rewarded.
Want to be notified of updates for delegators when Streamflow is in testnet and getting ready to go live? Sign up for updates here.
The role that transcoders are currently playing in Livepeer will be renamed to Orchestrators. If you’re interested in running an Orchestrator/Transcoder setup in Streamflow, or you would like to receive updates on this process, sign up for updates here.
Behind this naming change however, on the surface the job remains the same. Orchestrator nodes will continue to compete for transcoding jobs, but they will be able to load balance those jobs across their own pool of transcoder processes, or even to a CPU or GPU mining cluster that they are running.
Since job negotiation moves off chain, Orchestrators will have to provide better performance, reliability, and proximity to broadcasters in order to compete for and win jobs, and continue to perform them reliably.
There will also be increased competition, as the number of Orchestrator nodes on the network will increase well beyond the artificial limitation of 15 during the alpha.
Streamflow brings with it a number of opportunities for building on top of or contributing to Livepeer. All development, as usual, continues to be open source and Livepeer welcomes collaboration. There is also a nascent grants program being run by a community node that is beginning to provide opportunities for individuals and small teams looking to take on some of the suggested opportunities. The first round of proposed items include delegation tools, a community information hub, alerting applications, and translation.
- See this post for a list of the current grant opportunities.
- Check out the Livepeer Github.
- Drop into the Livepeer developer chat.
— — — — — — -
In summary, we’re excited to announce the proposal for the next major network upgrade to Livepeer — Streamflow. There continue to be open research areas, and community input and review is welcome to evolve these ideas into the first testnet implementation, and eventual network upgrade candidate. Work is already well underway on micropayments and offchain negotiation, and we’re hopeful that prior to the end of 2018 we’ll be able to demonstrate the cost savings and reliability on a testnet. Review, final implementation, code audits, and eventual production deployment will follow in the months to come.
Note: If you have questions or want to learn more, add your questions to the Reddit AMA Thread and join us Thursday at 12 pm EST to hear a Q&A with the community. Watch the call streaming through Livepeer’s infrastructure.