Discussion on Lowering the Number of Nodes

More on a better Node reward model:

Not including current airdrops node rewards are the biggest outflow of PRE from the treasury which as we all know is a huge risk for the future of the project and potential investors. So much so that a decision had to be made to increase the PRE MAX Token Supply from 500M to 1B and along with that Supply CAP increase, a new mint of an additional 250M PRE was required to sustain the reward outflows or otherwise risk destruction of the project. At around the same time drastic reductions in search and node rewards were made because of this same obvious problem. These reductions were needed to bring outflows down.

Everyone knows I am not a fan of Kucoin payouts and the airdrops because everything that was cut to reduce outflows has been negated at least for the duration of the anticipated airdrops. The bottom line is that High PRE outflows from treasury are a risk and bad… No or low inflows to the Treasury are also bad… both at the same time makes for horrible tokenomics and mass inflation causing all tokens to be worth less.

If you read anything above over the last year you can see I have not been against cutting costs and reducing outflows I simply wanted a more logical and better way to ensure cuts were done fairly and in accordance with the needs of the network.

The primary issue with the current reward model is the more nodes online with the same size reward pool reduces rewards for everyone even to extremes; a positive side of this simple model is a lower amount of nodes with same reward pool increases rewards for all. Hence Tim’s nodes coming offline have been the only reason rewards have not fallen drastically lower for everyone. But this isn’t the only issue, another is that the stake amount is half the reward although this is a great feature it should be more like 25% of the node reward pool. This leads to the last of the issues I have with the current reward model which is the 50% of rewards we get for running a node and actually servicing queries (the core utility of the “decentralized” network) is way to low at least with the current number of nodes. And this does not clearly reward performance or decentralization, a few things the network might be concerned about. Currently half of your node rewards is based on how much you have staked (more the better) compared to the total staked and the other half is basically if your node is online. Presearch can and should do better.

The result of no change in the current node reward model is that competitor with super low VPS costs, bad actors, or even enthusiastic new users wanting to run nodes could spin up thousands of new unnecessary nodes and since there is not a mechanism to prevent extraneous node creation all nodes would continue to get less and less rewards effectively to the destruction of the network. Any rewards beyond what the network requires is wasted outflows, since we know 60k nodes can easily handle 4-8M searches we have way too many nodes. I see this as a major risk and flaw of the system since bad actors and/or exuberance could result in ALL NODE runners reduction of rewards (payouts could drop potentially well below node costs for everyone). This flaw has other ramifications the obvious result is a flight to cheaper and cheaper nodes, possibly in terms of VPS providers becoming more not less centralized into just a handful of VPS providers and decreasing network performance. These are all risks and attack vectors that need to be remedied.

This is my suggestion
Create a node reward system that takes the following factors into consideration:

Quality Reliability score for now should play a bigger role
Stake gives users some say in what nodes they want higher rewarded but if min stake is on this should be less of a factor than quality
loyalty/ service time If you have been running nodes longer and all other things are equal you should be prioritized for rewards
Vps provider percentage Jurisdiction of providers and number of nodes / provider should be considered because these become attack vectors or areas of risk to the project. more decentralized the better especially if quality is the same.
Location/ Jurisdiction of VPS or node (country and state) More counties and even down to the state level of decentralization the better
Number of nodes per user Should be limited at least for now as the project cuts rewards. This ensures broader spread and decentralization of the nodes and less risk of a bad actors or any one entity controlling too many nodes DECENTRALIZATION.

The following is a visual representation of the above reward factors

This is a hypothetical visual result of node ownership restrictions to prevent any 1 person or entity from controlling too much of the decentralized node network

Lastly if adopted we could easily cut the PRE outflows from nodes in half and still have a stable, performance focused, decentralized node network.

How would this be implemented: The total node reward pool would be split into three pools
-Staked amount (individual stake vs total staked PRE) 25% of the pool

-Core nodes fully rewarded (depending on the search volume this would ensure the best performing, most decentralized, longest service/loyalty time core required nodes for network would be fully rewarded in accordance with the needs of the network) 60% of the total node reward pool

-Buffer nodes scaled rewards (This is all other staked online nodes outside of the core pool they would be ranked using the same algorithm as above but this smaller pool would then scale these rewards across all the nodes in this pool the best getting as close to fully rewarded and worst getting next to nothing) 15% of the pool

Having the 3 separate pools gives max flexibility to all users while ensuring flaws in the current system are resolved and rewards are paid in alignment with the needs of the network. Some in the buffer pool with home nodes or super low cost VPSs may still be able to run nodes at a profit especially if they have large stakes on those nodes. The core nodes would fluctuate any nodes in the buffer that outperform or offer more decentralization would swap with lowest nodes in the fully rewarded pool. This design should be more efficient and effective creating a sound foundation to scale from. As growth, searches, and increased income occurs the total pool size along with the requisite number of core nodes would grow.

Below is hypothetical scenarios with some variable changes.

Low PRE pice still preserves a reasonable reward for core nodes needed for the network. $1 per month with a standard 4k stake + core node reward.

Increase in PRE price means better rewards for all.

As you can see the total number of staked PRE vs your staked amount has an effect on your rewards. Every node would get the staked amount + either full core reward or a scaled reward. Numbers for scaled rewards is an average based on the number of buffer nodes my ideal would be best node in that group would get the highest amount but would start slightly below the fully rewarded amount and the worst would get next to nothing. This disincentivizes poor performance, and lack of decentralization, reducing unnecessary redundancy. However the entire buffer network is fully operable capacity to actively service network requirements.

1 Like