Farming rewards proposal: adjusted handling of violations

#1

Hi all,

I would like to propose a slight adjustment to the updated rules regarding farming rewards. As of last month there are new rules/requirements for farming in order to receive the rewards; 95% uptime over the entire month, or at most 1 farmerbot violation. In general I am all for rules that incentivise farmers to deliver quality/stability and/or punish those that fail to deliver. I did vote ‘yes’ to this proposal myself.

However, the current design is not well thought out in my opinion. In case of a violation (a lot easier to detect now, thank you scott!) of >5% downtime (regular) or 2 farmerbot violations, it would mean that the farmer does not get rewards for the month, that is not even finished yet. He now has 3 options:

-continue to deliver as best as possible and accept the loss of tokens
-shut down for the rest of the month and start back up next month
-wipe the node and reregister in order to start fresh

The rational options would be 2 or 3, but both are undesired outcomes. Let’s say the farmer has several deployments on his node, but had a power- or network outage on the node’s location. He can only get to this location 2 days later in order to fix this. Currently, contracts are not cancelled when a node goes down, and restarted when the node restarts. But the farmer would now be rewarded for choosing not to restart the node (in terms of saving on electricity, because no tokens will be earned regardless) and he would also be rewarded for deleting the node and all its deployments (because then he can start earning again rather than having to wait for the end of the month).
The same applies to farmerbot users, they are rewarded for wiping and reregistering in order to at least get some tokens for the month.

I believe the design should be such that, at all times, it is beneficial for the farmer to continue delivering the best quality/stabilty as possible, without having to mess with new node id’s etc.

Hence my proposal:

In case of a violation (either >5% downtime or 2 farmerbot violations), the farmer will lose rewards up until that day instead of for the entire month. The counters (downtime/violations) will be reset to zero.

This adjusted punishment should be enough to fulfil its goal as to incentivise the farmer to avoid any violations, and it remains beneficial to just fix the issue asap, without having any benefit to wipe and reregister.

I cannot think of an argument against this. Please consider it :slight_smile:

#2

As it comes to 95% of always on servers I don’t agree with you, I also want more serious penalty because when a server is rented this is work and threefold must have the same professionalism as other professional competitors like azure, google etc. Even 95% uptime is low enough and from now on that the farmer is getting 50% of rewards from the rent, I think there must be also collateral that will loose if the node is not up on a specific timeframe no matter the 95%.
But for the farmerbot controlled nodes I agree with you, because it is just a backup CPU and capacity, so it shouldn’t be so strict, I believe we must give more slack to backup solutions and be more strict on the rented nodes.

#3

I also like the idea of collateral. The point i was mainly trying to make, is that punishments should be based on past performance and not on possible future performance, because that will always lead to ways of tricking the system.

#4

If I am not mistaken if you don’t meet the 95% up time you still get tokens but not the full amount. You will always be getting something for being online. You will not be getting nothing for being online. That something will be based on your up time. The more the up time the more tokens. The less the up time, the less the tokens.

#5

Hi all, reading this thread as a newbie to all of this and not knowing how it all works yet, but wouldn’t it be simple if your reward is equal to the % up time? So 100% up time = 100% reward of your configuration. 66% up time = 66% of you max reward for your configuration?
I think we should not talk in terms of punishment which sets a negative vibe to what is being pursued. Instead let’s only talk in rewards and create that positive vibe we all need.

#6

My understanding is an uptime of 94.9% results in that node earning zero tokens for the minting period in question. A node is allowed a little over 36 hours of downtime per period, and each period is a standardized block of 30.45 days (730.8 hours).

Current state of rewards is outlined here: GEP - Minting and Farming Rewards Updates (March 2024)