Validators minimum wage

Governance
, , , ,
#1

I am not Personally a validator however I know the importance of them, to go as far to say that they are pretty much the back bone of a secure, efficient blockchain so I am all for setting a minimum wage in place but here are my thoughts.

Personally I feel some kind of structure needs to be put in place to prevent abuse of a minimum wage. We should set it so only the top 100 are eligible for this, more than 100 validators are not needed in my opinion and will simply dilute the minimum wage too much.

The minimum wage should also be given based on up time etc, if said validator was down 30% of the month then they should only receive 60% and the other 10% be given to those with 95%+ up keep, this rewards those who are benefiting the network more and will hopefully result in more strength. (Another option is to put the 10% penalty so to speak into the reserve)

The minimum wage should only kick in when ICX price hits a certain level where x amount of validators start to face losses and be turned off when ICX price hits 100% above this price point (Done quarterly due to price volatility)

What I would like to see in return and I am sure others would agree, in return for the network supporting validators in these times, when the time comes it would be nice to see the same level of support back because it may be a struggle now but if ICX price hits a level it can be the complete opposite in terms of profitability. I think once we hit a certain point where all validators are making nice profits, they should all dedicate a percentage of profits to offset carbon and more, so we become a carbon positive chain. This won’t cost much due to the chain already being pretty carbon friendly compared to other chains and it will be a great look for ICON in general, as years go by being a carbon positive chain will look much more appealing.

How to get funds for minimum wage, this question is hard to answer due to price volatility, if a percentage is set from fees or inflation that $ amount can change drastically each month. Lets say operating costs are $800 a month, thats $80,000 for 100 validators, I feel either a grant should given from the foundation or a CPS proposal should ask for $160,000, this will give the network a 2 month reserve for when the minimum wage kicks in, while this reserve is in place we should also start taking 0.5% from staking rewards and an equal amount from CPS inflation, if ICX hits 20cents this will be generating around $80,000 a month and needs to be sold instantly to a stable coin, the 2 months reserve should give enough head start to build up a nice reserve, considering some ICX will be sold for a much higher price (at 40 cent this will give 2 months worth of minimum wage funding).

We should not take the entire amount from the staking rewards, this will upset investors but if a grant is given from the foundation for a 2 month reserve it shows goodwill from the foundation and if the other 0.5% is coming from CPS inflation then stakers will be more willing to part 0.5% considering the circumstances and importance (On the basis that this will not be fixed and taken away when validators become profitable again).

It is annoying that there are so many dead votes in the network, perhaps stop paying staking rewards to those and force people to re-allocate their votes.

3 Likes
#2

Interesting thoughts. Have you discussed with anyone about possibly submitting a Network Proposal, which would modify the ICON Rewards & Penalties system somewhat based on these ideas?

I think if you get enough support it would be submitted, voted upon, and perhaps accepted.

It may be a good idea to move this discussion into the Governance category

1 Like
#3

This is a good discussion. Minimum wage incentivizes more validators to operate several nodes at a loss.

It’d be great to hear more validators make proposals on this topic. Voter concentration is a broader DPoC problem. I understand this is a difficult problem to solve.

1 Like
#4

Some thoughts on this from the perspective of a validator. Please note that I am quite jet lagged at the moment and these aren’t really fully formed ideas at the moment. Just wanted to get some of my thoughts down to kick off the discussion.

First, I want to establish my position on what a validator should expect from the ICON network. I’m not sure if this is still the case, but when we first deployed nodes, the recommended spec was a fairly high-end server on AWS. Following the migration from ICON 1.0 to ICON 2.0, the blockchain database format was changed to a format that requires more storage space. Currently, the blockchain database is ~1.2 TB in size. At RHIZOME, we’re currently running the recommended spec and pay ~$1,000/month.

Despite the $1,000/month hard cost of running a node, I think it’s important for validators to have some additional compensation for other node-related maintenance work such as keeping up with software updates, monitoring and scaling disk usage over time, and essentially being always on call for network emergencies. Thus, from my perspective, the minimum cost to run an ideal node operation is $2,000/month.

I use the word “ideal” because it’s really what I consider to be ideal. Over the past year, we’ve seen a number of public discussions from validators that are openly discussing how to reduce node costs. Frankly speaking, these sorts of discussions make me cringe. For example, validators have talked about moving their nodes to lower-tier hosting providers like Hetzner where it’s possible to set up a node for $200/month or even less. In fact, my $1,000/month figure has even been called out by one validator as a disingenuous way to justify more node income – which is just stupid IMO.

This sort of stuff makes me cringe because validators should not be bragging about who has the lowest node costs, and validators who shell out the big bucks for a high-quality providers should not have to feel like they’re overpaying – this is bad for ICON both internally and externally. Internally, it suggests that some validators have little knowledge about server providers, networking, redundancy configuration, etc. Externally, it doesn’t match up with the “enterprise-ready” narrative, which seems to be one of ICON’s general focuses. For example, if I were an enterprise looking to deploy my application on a public blockchain, why would I choose the network where validators are openly talking about who can run a node for the lowest cost possible? I would want to choose the network where I know validators are running on high-quality providers like AWS, GCP, Azure, etc. This is especially true for a network like ICON where blocks are produced by the top 25 validators. In summary, I think block-producing validators should be paid at least $2,000/month for running a high-quality node operation.

With ICX at $0.35, the #25 team is making $1,575/month which is less than the $2,000/month figure. So from my perspective, the network is already underpaying validators and something should be done in the short to medium term to reduce the chance of validators scaling down to lower specs or closing up shop altogether.

Now the question is what should be done… and that’s a tricky question.

I don’t think a minimum wage for validators is feasible because the consequences of the answers to the questions below are murky at best:

  • Do all validators receive the minimum wage, or only #1-25 (block producers)?
  • Do validators making more than $2,000/month (let’s assume that is the break-even cost that the network agrees upon) also receive the minimum wage?
  • Do validators need a minimum delegation amount to receive a minimum wage?
  • Do validators have to post a full bond in order to receive the minimum wage?

Do all validators receive the minimum wage, or only #1-25 (block producers)?

Restricting minimum wage to validators ranked #1-25 would probably piss #26-100 off because people are emotional. Realistically, #26-100 do not need the production spec because their nodes don’t produce blocks – they only sync. Despite that, they would probably still try to come up with some justification as to why they also deserve the minimum wage because everyone loves money.

Do validators making more than $2,000/month (let’s assume that is the break-even cost that the network agrees upon) also receive the minimum wage?

In modern society, the concept of a minimum wage exists to protect people at the bottom of the economic ladder. In other words, it’s not something that is given as an additional perk to every employed person. I’m not too read up on how minimum wage is determined in society, but it’s probably a number that is agreed upon by government representatives who don’t really receive a direct economic benefit from the minimum wage. In a decentralized network like ICON, economic parameters like ICX inflation and something like minimum wage would be agreed upon by validators. Unlike minimum wage in the “real world”, a minimum wage on ICON could ALSO directly benefit the validators. So in order to make the most profit possible, validators would be incentivized to set as high a minimum wage as possible. For example, instead of $2,000/month, validators may agree that $5,000 is a “better figure” for X,Y, Z reasons.

Do validators need a minimum delegation amount to receive a minimum wage?

Without a minimum delegation amount, a single entity could simply pay 2,000 ICX multiple times to set up multiple nodes to farm the minimum wage. With a reasonable minimum delegation amount, new validator teams (some of which may end up being quality contributors to the ecosystem) would be put at a disadvantage.

Do validators have to post a full bond in order to receive the minimum wage?

Same issue as above.

For these reasons, I don’t think a minimum wage would necessarily work.

Looking at the current validator rankings, I think the more pressing issue is distribution of votes rather than the amount of votes each validator has. Validators #58-151 have not posted a bond (literally 0%), yet they still have ICX delegated to them. Some of these are probably self-delegated teams, but most are not. Before exploring something like a minimum wage, I think it may be better to address the ICX delegated to unproductive teams. I call them unproductive because they literally serve no purpose beyond taking up votes that could be allocated to a productive team. By not posting a bond, these teams have also signaled that they don’t intend to contribute to the ICON ecosystem. Like @minhx said, this is a broader DPoC problem.

I’d be interested to see if we can come up with some mechanisms to increase vote allocation to productive validators (those that have at least 0%) bond.

Maybe something like a staking reward decay for ICX that is delegated to validators that do not post a bond – meaning at some point, reward for delegation to validators with 0% bond drifts to zero over time. Even though this might frustrate some voters, I think it’s very reasonable. At the end of the day, the network is maintained by validators, and validators are put into power by voters. Thus, voters need to do their job and select validators that are interested in participating in the game. If voters don’t do their job well, reducing staking reward is reasonable.

Another idea is adding an additional inflation bucket for productive validators. Right now there’s I_prep, I_cps, I_voter, I_relay. Maybe add an I_something, which can be used to reward validators that participate in additional responsibilities beyond just being a registered node. For example, maybe part of the inflation to I_something goes to validators that participate in the CPS. I think this could be the “minimum wage” we are looking for because it’s a fair trade – validators give additional time to the ecosystem in exchange for money.

Getting sleepy so will leave it at that for now…

3 Likes
#5

Half awake myself so excuse any broken points
Fully agree that we don’t want validators to start cheaping out on the quality of operations, really doesn’t look good but I also think we should recalculate the recommended running costs for a last resort fall back if minimum wage fails, in a bear market we will not need the current running recommendation In my opinion as chain activity will fall drastically, so I am sure some changes could be made until market conditions pick up.

As for the $1000 wage figure you mentioned, I agree but on a different view. I don’t think the actual work itself is worth $1000 based on time required, as I have run nodes in the past and maintenance imo is low (Perhaps it requires more time than the ones I have run, if thats the case please excuse my ignorance on this). I do however believe $1000 wage is justifiable based on the importance of validators in general (Only saying this as I know others will try call out the wage claiming validators don’t require much work etc when in fact without them we basically have nothing)

Good points, I believe we should have two minimum wage categories as operating costs for the top 25 will be more than those for 26-100, we should calculate the optimal specs for validators from 26-100 and get them to all scale back to these specs, this will save cost straight away and we can set a reasonable minimum wage for these validators.

If validators are making over $2000 (and we all agree this to be the wage/cost) they should not receive the minimum wage, this is being put in place to support validators and to stop them from having losses, as you mentioned it is not something to be added to be just an additional perk/bonus. It should kick in when a validator falls below the $2000.
I see your point about validators having the final say due to voting, I guess with that we would have to hope validators do not abuse this (I would personally steer clear and move votes if I see validators do this, really not a good look but I have faith in our validators), the only other way is to go straight through the CPS system but this probably isn’t the best route for a few reason.

minimum delegation
2000icx will be fine as long as we cap it to top 100 validators and only make it accessible to those who have bonded, which as you mentioned this is currently the top 58. As to be completely honest most validators are not really contributing anything, I would much rather we focus on building strength with the validators that are contributing including finding a way to force a re-allocation of votes, we don’t need 100 validators, not yet anyway.

What to do?
We need to focus on compounding strength not numbers, if we loose some non-productive validators that is more than fine imo, it will then compound into strength to the main validators that are contributing. We need to focus on the productive validators, those who produce blocks and those bonded. If you were working a minimum wage job and being non-productive you would get fired, so I don’t see why we should pay such validators.

The minimum wage should be ready to accommodate the top 100 but on these requirements

Requirements to get minimum wage-

  1. Be in the top 100
  2. Be bonded
  3. Be falling under the wage/operation cost bracket we set (say $2000 for top 25 and $TBD for 26-100)
1 Like
#6

I’m moving this discussion to the Governance category and adding some tags

#7

@bwhli @WeMove thank you for your thoughts. I’ll add a couple quick things - but am still doing some more research before I provide anything substantial.

I would like to clarify here that the Top 22 validators are always block producers, while 3 randomly selected validators are chosen from the set of sub-validators ranked 23-100.

Since all top 100 nodes have the potential for block production, maybe only the top 22 absolutely need the production spec but the sub-validators don’t?

Let’s not forget that once the slashing penalty > 0%, it would be in the best interest of a node to run the optimal setup to avoid slashing.

I would say making a comparison to “real world” minimum wage is hard because an employer doesn’t need to ensure their employees are operating “above costs”, as the employee’s only cost is their time (opportunity cost).

I think maybe rephrasing “minimum wage” could help us think about this a little bit better. Network inflation towards block producers is essentially just the network paying for its security. Maybe thinking about it as “what amount is our network willing to subsidize to ensure its security”?

Something like “how much does it cost us for our Minimum Viable Security” and “how much would it cost us for our Optimal Network Security” could be good things to think about.

Whether top validators should receive this subsidy or not, Validators outside the Top 100 should not receive any subsidy as they have no potential for block production in our current system and they do not factor into the security of our network (since that’s what we’re essentially paying for).

I agree with this point. Most common across the industry, the purpose of Validators is 1) Block Production 2) Network Governance. If a “validator” is doing neither of these things, they should not be eligible to receive rewards.

If we were to reallocate the buckets to 90% network inflation towards Validators (10% to CPS) and Validators set their own commission rates (distributing the rest of the rewards to delegators), then those who are voting for “unproductive” validators would receive 0 staking rewards and it would be their responsibility (and in their best interest) to redelegate to “productive” validators.

This is how the reward structure is built for the Cosmos Hub (plus a couple other minor things).

My personal opinion is that CPS validator rewards should not be a fixed amount.

I’ve been looking more into incentives for CPS participants, but have nothing concrete to share yet as proper DAO incentivization is quite a large & difficult issue across the industry right now.

Overall, lets continue this discussion!

#8

Network useage should define the node that is ran.

You cannot want to ensure a minimum wage based upon a spec that is bigger than our shoes right now. I would much rather this minimum wage goes towards more building that improves token price which then inevitably pays the ‘minimum wage’ for high end nodes.

To ensure minimum wage, based upon bad token price action, is a back-to-front way of doing things imho. What will you end up doing? inflating the token for node operators to run high end nodes, that way outpace demand?

#9

Last time I saw a discussion about reducing icx inflation. Guaranteed minimum wage to verifiers How about keeping icx max inflation below 3% by reducing staking rewards? I think that 4-5% of the staking reward is sufficient. I know about the staking rewards of Ethereum like this

1 Like