Time to talk about the DATA token economics of the Streamr Network! The token economics will be released in the 2022 milestone (Tatum), and actual implementation work will start as soon as this year’s milestone (Brubeck) has been shipped. Although the current state of research and planning are at a pretty advanced stage, please bear in mind that what you see here may still be subject to change. Before diving into the roles and mechanisms involved, let’s briefly discuss why token economics are needed in the first place!
Why the Streamr Network needs token economics
In the Streamr Network, both data publishers and subscribers are nodes in a peer-to-peer network. Nodes involved in a stream of data connect to each other in a certain way – forming the stream’s topology – to perform the main function of the Streamr Network: the dissemination of messages published to that stream.
So every node that joins a stream both consumes data in that stream AND helps relay it onwards to other nodes interested in the stream. Since everyone who uses resources also contributes resources back to the network, can’t the system work without any token incentives, based on reciprocity alone?
Under perfect conditions perhaps, yes, but the real world is far from perfect. In practice, the reliability of message passing is reduced by the following
- Churn – Nodes joining and leaving the stream can add instability to the topology, which could disrupt message flow
- Free riders – In a decentralized network, there’s no way to enforce honest behaviour. Some nodes might selfishly consume messages but refuse to propagate them onwards
- Malicious nodes – Some nodes may intentionally try to disrupt the network
- Node failures – Nodes can crash, run out of memory, get overwhelmed with data, get disconnected, experience packet loss, etc.
So, very roughly, what determines the quality of service in a stream’s topology is the ratio of honest, stable nodes vs. nodes that perform questionably. If your use case needs honest, stable nodes to make your data flows more robust, how do you incentivise good nodes to join? Well, you pay them. This forms the basis of the token economics (tokenomics).
Traditional software products often have freemium and paid plans. Decentralized protocols usually have more of a grayscale spectrum as a result of being subject to true market conditions – you pay more for better service. Gas price on Ethereum is a good example: you bid a higher gas price to incentivise miners to execute your transaction faster. A similar economy is at play on Streamr: you pay less (or even nothing) if you’re happy with best-effort performance, while you can pay to incentivise nodes to make your stream more robust and secure.
For the sake of clarity, the Streamr Network tokenomics should not be confused with the transactions happening on the Streamr Marketplace. On the Network, people pay for infrastructure costs; for data delivery. On the Marketplace, people pay for access to data content. Here’s an analogue: when you order an item from an online store, someone (either you or the store) pays the postal service for package delivery. You also pay the online store for the item you ordered, or the package content. Note that you can very well use the Network for data delivery without using the Marketplace at all, just like you can send and receive packages without ordering stuff from online stores.
The roles in the Streamr tokenomics
Publishers and Subscribers are already familiar roles in the Streamr Network. These roles are only related to the data flows in the network, meaning these roles could be seen as being one layer ‘below’ the tokenomics. In the below diagrams, data flows are shown in blue.
The introduction of token economics defines three new roles: Sponsor, Broker, and Delegator. These roles use DATA tokens to participate in the incentive mechanisms, and those value flows are shown in the diagrams below in orange.
It should be noted that the roles are independent of each other and can be mixed and matched by all actors depending on their goals, for example, the same person could be a Sponsor, Publisher, and a Delegator.
A Publisher is simply a node through which certain data enters the network. The data usually originates in an adjacent application that interfaces with the Publisher node, with the goal of delivering that data to Subscribers via the network. Publisher nodes relay the messages to other nodes they are connected to.
A Subscriber is a node in the network that wants to receive messages from a stream. Just like Publishers, they also relay the messages to other nodes they are connected to. Subscribers may have a range of different motivations for joining a stream – there could be an adjacent application that wants the data, they could be Brokers who mine the stream by becoming Subscribers (see below), or they may even want to help the stream simply for charitable reasons.
Sponsors pay to secure the operation of a stream. They pay DATA tokens into a smart contract called a Bounty. Essentially, the Sponsor says “I want to spend X amount of DATA tokens over a period of time T to improve stream S”. The Bounty contract releases the funds over time to Brokers who are mining the Bounty (see below to learn about Brokers). It should be noted that a stream can have many Sponsors, and they could be anyone at all, including the Publisher(s), Subscribers, or a third party.
Broker nodes are the miners in the Streamr Network. They are nodes that constantly monitor which Bounties are available – Bounties are smart contracts on the public blockchain, so they are visible to everyone. Broker nodes choose the Bounties they want to start mining, stake DATA on them, and become Subscribers in the related streams. The promise of the Broker is: “I am an honest and stable node, and I’ll join the stream topologies to help stabilise and secure them”. Brokers don’t subscribe to a stream because they’re interested in the data, they join because they want to earn a share of the DATA tokens flowing through a Bounty. The Broker can claim their rewards periodically to withdraw earned tokens from the Bounty contract.
Brokers are expected to be honest, following the protocol rule of properly forwarding messages to other connected nodes. And they are also expected to be stable, with good uptime along with sufficient bandwidth and hardware resources to handle the traffic of the stream. If they fail to meet these standards, they could be kicked out of the Bounty and their stake could be slashed.
The amount of DATA tokens the Broker stakes on the Bounty determines the size of their share of the token flow. The Broker will generally stake on Bounties where they can earn the best available yield for their stake. Other Brokers in a Bounty are there to share the cake, so overcrowded Bounties may not offer the best yields, and some Brokers will decide to mine other Bounties instead. Like any open market, the market of servicing Bounties will always gravitate towards achieving equilibrium between supply and demand.
Delegators are DATA token holders who don’t want to do mining themselves, but would rather earn yield on their tokens by supplying liquidity to Brokers. In exchange, they earn a share of the Broker’s rewards. Since Brokers need to stake tokens to mine Bounties, having access to tokens from Delegators enables them to earn more from Bounties, creating a win-win situation.
Delegators select Brokers to stake on and deposit tokens into the Broker’s Stake Pool smart contract. The funds in the pool can then be used by the Broker for staking on Bounties.
Observability and free riding
While the above processes and roles may seem quite straightforward, one of the key challenges is preventing Brokers that don’t actually do the work (of joining the stream’s topology and relaying messages to connected peers) from earning Bounties.
Since the Brokers place a stake on the Bounties, they could be slashed for not doing the work. The problem is proving to the smart contract that a particular Broker did not do the work or otherwise live up to its promises. There’s limited observability in the network, as basically only the peers connected to a node really know what the node is up to. On the other hand, only the tracker knows which nodes are connected to each other. By combining attestations from a Broker’s peers as well as the tracker, it could be shown with reasonable confidence that a Broker did/didn’t fulfil the requirements of servicing the Bounty – potentially leading to slashing the Broker or at least kicking them out of the Bounty.
Then again, the peers or the tracker could be malicious and give false attestations in order to harm a particular node! As you probably realise by now, this is pretty complex. As is usually the case with decentralized systems, nothing is foolproof and every such system has some conditions under which it fails or becomes less reliable. Understanding those conditions well is the key to establishing what parameters work in the real world and what guarantees can be given about the system in practice. And indeed, much of the remaining work is exactly in this area.
As part of the rollout of the Brubeck milestone, we’ll also be releasing a new version of the Streamr website. Among other improvements, the updated website will contain a new token economics page, which will further illustrate the concepts and mechanisms described here. However, as I mentioned in the beginning of this blog post, there may still be some changes and tweaks to the plans. Once we start working on the Tatum milestone after the release of Brubeck, token economics will become front and centre in the team’s everyday work, and any loose ends will be ironed out.
What do you think about the current progress and plans? Join our Discord, especially the #data-tokenomics channel there, to chat more! Also make sure to participate in the upcoming Brubeck testnets to get hands-on experience in running a Streamr node. Not only will you set yourself up for building on Streamr, you can also earn a share of 2M DATA available to testnet participants!