RNDR Tokenomics Update: Multi-Tier Pricing (MTP) – Render Token
Dear RNDR Community,
More than a month into the RNDR Mainnet, we are excited and happy to see users rendering scenes on our network and getting RNDR miner rewards fulfilled in real tokens. As we work through adding additional features, we wanted to share with the community our current thinking on select aspects of the economics of the RNDR network and how it flows through various points on the platform.
A core question for the network is how the specific rendering power of a token will shift with changes in the market price of rendering. Naturally, when you have users buying and selling tokens in order to utilize the platform and render projects on the network, you will have a consistently changing underlying price of the token. This could be problematic for users that require rendering services on a consistent basis, or for miners looking to have a consistent and predictable inflow of RNDR for their work. Both artists and miners need to forecast costs to effectively use the network.
RNDR Multi-Tier Pricing
One solution that we have been working through internally to minimize this uncertainty for users and miners is a process that we call Multi-Tier Pricing (MTP). MTP allows users to select what Tier of power they want their job to be processed based on parameters like speed, system specs, and security. The pricing for each tier will be stable but periodically adjusted for changes in GPU performance, GPU cloud rendering costs, and supply / demand data. For example, we will make adjustments with new OctaneBench releases.
The Reason for Multi-Tier Pricing
When designing RNDR, an overarching economic goal is to clear as much GPU compute supply and demand as possible, maximizing overall use of the network. In economic terms, we want to eliminate deadweight loss — when supply and demand are mismatched, leading to unproductive waste. Monopoly or distorted pricing, the cause of deadweight loss, occurs when producers set prices artificially high to extract rent or when pricing does not effectively capture demand. Blockchain peer-to-peer marketplaces offer remarkable opportunities to reduce this type of waste and generate new forms of economic productivity. For RNDR, GPU cloud rendering needs to become democratized in order for next generation holographic media creation to be widely scalable. Reducing deadweight loss helps artists (who don’t forego cloud rendering due to high costs) and miners (who minimize idle GPU time).
RNDR Multi-tier pricing efficiently matches supply and demand by creating a set of preferences that users can choose from to best suit their rendering needs and price elasticity. Artists with inelastic preferences — for example, time pressures like a critical deadline or highly-complex scenes that can only be processed with specialized hardware — will be able to have their demand met using a higher tier of nodes with priority. Meanwhile price sensitive artists, who mainly want cheaper GPU cloud rendering options, will have the ability to get work done at highly competitive rates as nodes become available on the network. When both elastic and inelastic demand is met on the network, RNDR maximizes use on the network. Further, higher tiers reward the creation of premium services that some artists need like high performance specialized GPU clusters and more secure node configurations. With supply and demand efficiently clearing for maximized network usage, RNDR creates a sustainable multi-sided GPU rendering marketplace.
A Multi-Sided Network
In order for RNDR to sustainably grow, it needs to balance both artist and miner interests. For example, if pricing did not allow miners to pay their hard costs like electricity and infrastructure expenses, there would be no one to process rendering jobs for artists. Similarly, if rendering was too expensive or slow for artists, there would be no jobs on the network for miners to process. Keeping these interests in balance over the long term is the goal of the network’s economics.
There are a number of heuristics the team is working with in order to create a sustainable equilibrium. The first is that miners need to process rendering work above marginal costs, allowing them to pay electricity and operational costs with sustainable operating margin. As part of our marginal cost calculations, which we will use to help create a price floor, we are including GPU infrastructure (see Figure 3). While it is debatable whether these are sunk costs or not, and whether they should be considered marginal costs, we want the network to function sustainably over the long term, allowing miners to upgrade to the latest generation of GPUs. Because GPU upgrades create better supply and a more productive network for artists, we are including GPU hardware in marginal cost. On the artist side, we want RNDR to become price competitive with other forms of rendering so that unbiased cinematic GPU path-traced rendering becomes the industry standard. We want to remove trade-offs between speed, quality and resolution — providing artists the ability to scale to the cloud with reduced cost and time pressures. Below is an overview of how the network achieves these goals through MTP.
Multi-Tier Parameters
In this example, you can see the ascending Tiers as the relative cost and power for each Tier increases from Tier 3 → Tier 1. If you are an artist that doesn’t need the work to be completed quickly, you can choose a Tier 3 iteration of a job, which would cost less than a comparable Tier 2 and Tier 1 iteration, all things equal. We aim to provide a solution to the problem of different needs by giving users a choice to pay different amounts based on what they are looking to get out of their specific render job.
[A Note on Tier 1 OB Adjustment]: Tier 1 reflects the initial RNDR pricing based on OctaneRender Cloud (ORC) secure GPU cloud rendering costs. As you can see in Tier 1, we are adjusting the equation for 1 RNDR so that it can be easily normalized to $1 and an hour of OB work, helping artists easily calculate rendering costs. We are also updating the amount of work artists get for 1 RNDR to reflect the latest efficiencies in OctaneBench 4, which is about 10–15% faster on Pascal GPUs and 25–30% faster on Turing/Volta. The adjustments do not substantially change the price of RNDR, they are focused on updating RNDR for increased efficiency in OctaneBench 4, as well as intended to make it easier to calculate for artists — i.e. 1 RNDR = $.25 of rendering work at 25 x OB4*H and $1 of RNDR work =100 x OB4*H.
On the miner side, we will be implementing node specific factors for Tiering. Examples of the factors that we may potentially include are: Total System OB, Node History (Successful % of jobs processed), Minimums forVRAM and RAM. These figures will be processed and combined through a Composite Score system, similarly to our Reputation Score. We are currently working on releasing a User Manual that will explain this further, but wanted to share this with the community as a primer for the upcoming documentation. An example table to go along with Figure 1 is included below:
As we work to provide a way for miners to exchange tokens in order to cover basic energy costs, we have developed a tool to assist with the power consumption mechanics and costs associated with rendering. We are currently working to release this tool to the public; in the meantime, here is a simple example screenshot of the calculations and expected output of the figures. These figures help us calculate operational costs that need to be accounted for token pricing. For simplicity we assume 100% GPU utilization at full TDP and the global average electricity price of $.12 kWh. Additionally, we will be ensuring the tool accounts for higher electricity costs, so that miners in those areas are not operating at a loss.
This post is intended to preview some of the network economics that are designed to facilitate efficient peer-to-peer rendering. We will be releasing additional information in user manuals to support successive launches after we complete the Genesis Beta period.
Join the Rendering Revolution!
- By subscribing to email updates on our mailing list!
- By following us on Twitter!
- Through our community chat!
- By subscribing to our subreddit!
- Checking out our Telegram!
Published at Sat, 10 Aug 2019 00:37:26 +0000
Bitcoin Pic Of The Moment
✅ Marco Verch is a Professional Photographer and Speaker from Cologne. ? This image can be used under Creative Commons 2.0. Please link to the original photo and the license.
By marcoverch on 2017-11-27 01:48:39
