Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

LRC-3 (Block Reward Restructure) #10

Open
KeeJef opened this issue Oct 9, 2019 · 66 comments
Open

LRC-3 (Block Reward Restructure) #10

KeeJef opened this issue Oct 9, 2019 · 66 comments
Labels
ORC Loki Request for Comment

Comments

@KeeJef
Copy link
Collaborator

KeeJef commented Oct 9, 2019

Overview

I'm creating this issue to strictly define what options the Loki team and Loki foundation has raised in unison with, LIP-5 @CryptoFirefly raised this same issue in #9 but the choices presented were not exactly in line with those that are being proposed by Loki Foundation or Loki Project team right now.

With the introduction of LIP-5 Service Nodes will fully take over block creation which means Miners will no longer have a role in the Loki ecosystem. This naturally raises the question of what we should do with the 45% mining reward that currently goes to miners. The Loki team and Foundation has been discussing 2 distinct proposals.

Proposals

    1. The 45% mining reward is never created and the existing rewards (50% Service Node and 5% Loki Foundation) are re-balanced, so each block distributes 95% of rewards to Service Nodes and 5% to the Loki foundation. This change would result in an overall reduction of new Loki being created by 45%

Loki per block (9/10/2019) approx would be
14.736 for SNs, 0.736 for LF. (95% SN, 5% LF)

  1. The 45% mining reward is given to the Service Node network, this means each Service Node reward would increase 45% to replace mining. This would mean 95% of the rewards go to Service Nodes and 5% goes to the foundation. The emissions curve would remain the exact same with the same amount of new Loki being created as now.

Loki per Block would be approx (9/10/2019)
26.6 for SNs, 1.41 for LF (95% SN, 5% LF)

*Note: transaction fees will also be paid to Service Nodes

Other questions

If you have any additional suggestions as to how to restructure the block rewards then you are free to make them here aswell.

There is an additional bonus question, which is when should this block reward restructure occur? Should there be a hard cut over where miners are eliminated immediately when PoS is enabled? or should we continue paying miners until we are sure PoS is stable and working live on mainnet? If ambitious we could decide to commit to a decreasing schedule now which slowly reduces the miners rewards overtime, reallocating that reward to Service Node operators.

Please write your feedback as a comment below, if you support either of the major proposals please signal that as such.

@KeeJef KeeJef added the ORC Loki Request for Comment label Oct 9, 2019
@pailakapo
Copy link

pailakapo commented Oct 9, 2019

Reduce emissions to 20 per block as follows:
On switch to PoS until proven stable:

  1. 95% of 20 to SN
  2. 5% of 20 to LF
  3. 8 loki to miners
    When stable:
    Remove miner reward

***EDIT
Since there is not a plan to change the emission schedule outside of #1 or #2, I definitely think #2 is the best option.

@CryptoFirefly
Copy link

@KeeJef
Could we do the following and probably edit / add to your original message, as most would read that.

  1. Add the absolute Loki per block. Like in
    Option one the Loki per block would be 14.736
    14 for SNs, 0.736 for LF. (95% SN, 5% LF)
    Option two the Loki per Block would be 28
    26.6 for SNs, 1.4 for LF (95% SN, 5% LF)

In either case the 95:5 is sort of the likely direction. What's open between the two options and a possible 3rd option, is the absolute Loki per block.

Option three would be, a plethora of options like to generate (whole number) 16, 18, 20, 22, 24, 26 Loki per block and distribute that across SN & LF.

I feel the ideal decision is something from Option 3, with the right modeling done for rewards in future. I like 18 or 20 Loki per block, option the most.

On mining rewards, I prefer 2 step process,

  1. Implement the big drop to say 20 Loki per block and within that have 70:5:25 ratio SN:LF:Mine
  2. Next fork, 95:5:0 - End of mining.

@KeeJef
Copy link
Collaborator Author

KeeJef commented Oct 9, 2019

I can put absolute amounts in, but obviously the block reward is still constantly decreasing, so they can only be considered accurate for a short period of time.

Regarding 3rd options I really don't think its a good idea to design an inherently new emissions schedule, there was alot of work put into the current assumptions we have made with with two papers being published on the cryptoeconomics.

I would prefer to keep the possible changes as simple as possible and not make major alterations to the block reward.

@CryptoFirefly
Copy link

CryptoFirefly commented Oct 9, 2019

True, I hear you on the work put in the cryptoeconomic papers :)

On the update above, i was suggesting more from final block emission at those options, as we are tending towards them. And those no.s i put are ~final figures.
Option 1 would halt at 14.736842105 Loki per Block (to maintain 95:5 or 14 : 0.736842105) &
Option 2 at 28 Loki per Block _(like today, with 26.6 : 1.4 ratio)

Having read all the point of views and i understand that its best to not change the emission curve too often and keep it stable, i prefer Option 2, to keep it 28 Loki per Block and 95:5 ratio.

+ 1 for Option 2

@Lucifer1903
Copy link

Lucifer1903 commented Oct 9, 2019

Okay so I went back and re-read the cryptoeconomics paper and decided to make the charts for the 2 proposals mentioned and 2 additional proposals that are a compromise between 1 and 2. Please see here #12

I would personally choose proposal 4 but I'm not opposed to proposals 2 or 3.

After re-reading the crypto economics paper I am very much against proposal 1 as it goes against the goals stated in the paper.

Edit: I'm also for the ambitious option of committing to a decreasing schedule now which slowly reduces the miners rewards overtime, reallocating that reward to Service Node operators.

@Skelaton4
Copy link

Skelaton4 commented Oct 9, 2019

I am withdrawing my opinion to gather more information.

@KeeJef
Copy link
Collaborator Author

KeeJef commented Oct 10, 2019

@Lucifer1903 what your analysis fails to account for is how a decrease in new Loki being created affects the real world USD price of Loki. If we assume that demand for Loki is fixed and the newly minted Supply is lowered as in option 1 we would assume an increase in the USD price of Loki. This has the knockon effect of increasing the real world ROI of Service Nodes, which is arguably more important than the Loki ROI.

Note too that with any increase in the USD ROI of we should see the amount of people willing to run Service Nodes increase (to a point), as the profits are now increased.

@alex76888
Copy link

My vote would be for option 1.

@Lucifer1903
Copy link

Lucifer1903 commented Oct 10, 2019

@kee yes that is arguably true, yet it is difficult to predict what the price of loki would be, that is why I copied the assumptions from the cryptoeconomics you guys released and why I'm for a middle of the road option.

As you said "there was alot of work put into the current assumptions we have made with with two papers being published on the cryptoeconomics" and I think people should go back and read those cryptoeconomics papers because it would be a shame to ignore the conclusions that were came to there.

Edit: maybe we need another cryptoeconomics paper if you think the conclusions in the originals are no longer relevant.

Edit: sorry I have to challenge this argument "If we assume that demand for Loki is fixed and the newly minted Supply is lowered as in option 1 we would assume an increase in the USD price of Loki. This has the knockon effect of increasing the real world ROI of Service Nodes, which is arguably more important than the Loki ROI.

Note too that with any increase in the USD ROI of we should see the amount of people willing to run Service Nodes increase (to a point), as the profits are now increased."

Even if the there is an increase in USD value of Loki it would only increase the value of loki held by individuals, it would not increase USD ROI or Loki ROI and it wouldn't increase willingness for new people to run service nodes because they would have to buy in at a higher price with less ROI in both Loki and USD.

You could argue that it would increase USD ROI if you look at early investors and compare their original investment with USD ROI but if we look at the actual value of their stake with the USD ROI then it it would be lower under option 1, if a better ROI can be found elsewhere they might as well sell their stake at the higher price and get the better ROI.

I hope that came across clearly, I'm not the best at putting my words together. In summary option 1 prioritizes the value of Loki over the USD and Loki ROI for service nodes, giving less incentive for for newcomers to run service nodes and giving more incentive for current service node operators to sell.

@Lucifer1903
Copy link

Lucifer1903 commented Oct 10, 2019

I corrected an error on the original proposal 1 chart (changed block reward from 15.4 to 14.7).

Loki_emissions-1-1
Loki_emissions-2

@kee I've made a new chat for the value of capital staked in service nodes + profits, instead of using the prices assumptions from the economics paper I have assumed that the price is unknown but will vary between each proposal by the amount of inflation.

Loki_capital_value-1

@KeeJef
Copy link
Collaborator Author

KeeJef commented Oct 11, 2019

Even if the there is an increase in USD value of Loki it would only increase the value of loki held by individuals, it would not increase USD ROI or Loki ROI and it wouldn't increase willingness for new people to run service nodes because they would have to buy in at a higher price with less ROI in both Loki and USD.

An increase in the USD value of Loki leads to a direct increase in the USD ROI of Service Node operators, since the Loki that they will earn from Service Node operation is now worth more in USD value. My contention is that USD ROI here is more of a motivating factor to new entrants to the Service Node network than the barrier of entry being raised (To a point, in my estimation about 55%-60% of locked supply)

@Lucifer1903
Copy link

@kee that's a fair assessment if we assume the main contention is the USD return for work put in to running a service node.

However I think running a service not isn't a contention because following the set up instructions is pretty easy and after that it's not difficult to passively monitor them with the LokiSNBot. In my opinion what it really import to get more service nodes is USD % ROI for the amount of USD capital that is staked at any given moment. Its the % that's important. Even if proposal 1 is going to give operators a slightly higher USD return for the work put in, I think most potential operators would be looking at the USD % ROI for their initial capital investment. We cannot know what the price will be in the future but we can assume that all other things being equal inflation will affect price by the by the amount that it is i.e. Let's say by year 3 all other conditions would cause loki to go to $10 (I'm using this because its a nice round number). Proposal 1 year 3 inflation is 6.68% so that would mean loki would be worth $9.33. Proposal 2 year 3 inflation is 12.72% so loki would be worth $8.73. Proposal 3 is 9.09%, price would be $9.09.
So yes re reduced inflation of proposal 1 would give a higher price but we can't predict the other conditions that would cause price change. So any potential new operator can only consider the the amount of USD invested into loki Stake, then the amount of rewards, then the effects of inflation on both the initial capital investment and the rewards. I've already done this in the last chat I posted and it can be seen that under proposal 1 even though the price of loki would be higher the total value of capita investment plus rewards is the lowest out of all options.

There's one more thing that I just thought of that I think is something that should be considered. The premium services that will have to be paid for by burning loki, I understand the prices for these are going to be set in equivalent USD, so If it costs equivalent $10 for a service then more loki will be burnt when the price is lower. This means that even though higher inflation would mean a lower price, it also means more loki would be burnt for the same amount of services. Under the higher emissions proposals inflation would be offset by a higher degree by burning. I don't know exactly how this would effect inflation and price, I guess the only thing that can be stated is that higher predictable inflation also comes with higher unpredictable deflation.

@shishi8565
Copy link

I vote for option 1. Makes the most amount of sense if you're holding a LT view.

@Cactii1
Copy link

Cactii1 commented Oct 16, 2019

You guys finally decided to go for the kill?

@jagerman
Copy link
Member

Let's say by year 3 all other conditions would cause loki to go to $10 (I'm using this because its a nice round number). Proposal 1 year 3 inflation is 6.68% so that would mean loki would be worth $9.33. Proposal 2 year 3 inflation is 12.72% so loki would be worth $8.73. Proposal 3 is 9.09%, price would be $9.09.

Just a clarification: you're conflating coin inflation with price inflation here, but they are two entirely different (and only loosely related) concepts. Basically the only way that the two inflations can be equal is if the market cap of loki is perfectly fixed and that seems hard to believe to me.

I believe that the determinant of Loki's future is far more dependent on building demand for Loki and that, compared to increasing demand, the two supply options being considered here are basically going to be a rounding error.

What it will affect, however, is the number of SNs operating on the network, and I think that is where we should be focusing.

Will 14/block give enough return for the number of SNs we want? Does 28/block? Something in the middle (for example: we could just jump off the block reward formula to a fixed reward of 20 and just remain there forever).

Changing the emission curve is something that should not be undertaken lightly: a lot of decisions have already been made knowing that we have a 28 LOKI tail emission. We've already changed it twice before; both decisions were firmly rooted in the network itself: first that the original curve tail emission could not sustain the SN network; and second that we wanted to avoid a problem caused by the first curve's long run 10k -> 15k ramp-up when introducing infinite staking.

I think if we want to change away from the current emission formula here the decision similarly needs to be based on technical consideration for the network and not solely on the effect such a change would have on the price of LOKI (in part because such a change is nearly impossible to predict).

@trekman7
Copy link

+1 for proposal 2 with the ambitious option (decreasing schedule of miner rewards)

@Azazel020
Copy link

I would go for option 1 but prefer to have a 45% to the SN's and a 10% to the LOKI Foundation.
This 5% extra needs to be taken from the miner rewards so 45% of the rewards will not be created.
Dash also works with a 10% and is a good example of how it could work.
I think that's necessary to secure future developments.

If we go for that 10% then the LF shouldn't run a ''10% foundation nodes''.
Instead of those nodes the foundation should use the Foundation funds to stake on pooled nodes that have a max fee of ….% .
This will encourage more people to set-up a node.

Not sure if it fits in this discussion but maybe it's also an option to give the nodes a different reward % based on their geolocation.
Not sure if this would be technical possible but it gives the incentive to a SN Operator to use other provides and create diversity in the geolocation of the nodes.

@VINCEG-afk
Copy link

I am biased towards option 2. Don't mess around with the social contract. This is the emission which was agreed to and the whole basis for your crypto economics paper. Risk is however economic incentive if LOKI price goes too low and it become not viable to run node. So my suggestion is option 2 with staking requirement increase.

@Lucifer1903
Copy link

Lucifer1903 commented Oct 16, 2019

@jagerman, I agree with everything you wrote but I think you misunderstood my earlier comment (I'm not the best at explaining sometimes)

What I was trying to convey is that coin inflation does have an effect of price inflation, to what extent is debatable but from what I understand higher coin inflation causes more price inflation and lower coin inflation causes less price inflation (though building demand for loki via use cases is going to be the determinant factor of loki's future and price).
I main point I was trying to show is that even if lower coin inflation causes less price inflation under proposal 1 the USD profits of service nodes would be lower than the other proposals (and I would think less USD profits = less incentive to run service nodes).

Maybe I am mistaken but I just want to make sure that what I'm saying is understood correctly so that it can also be debated correctly.

I agree with you on everything else you wrote.

@will-miningstore
Copy link

I would go for option 1 but prefer to have a 45% to the SN's and a 10% to the LOKI Foundation.
This 5% extra needs to be taken from the miner rewards so 45% of the rewards will not be created.
Dash also works with a 10% and is a good example of how it could work.
I think that's necessary to secure future developments.

If we go for that 10% then the LF shouldn't run a ''10% foundation nodes''.
Instead of those nodes the foundation should use the Foundation funds to stake on pooled nodes that have a max fee of ….% .
This will encourage more people to set-up a node.

Not sure if it fits in this discussion but maybe it's also an option to give the nodes a different reward % based on their geolocation.
Not sure if this would be technical possible but it gives the incentive to a SN Operator to use other provides and create diversity in the geolocation of the nodes.

I think an incentive based on geographic location is paramount, I am not sure how this would be made possible, and I am also not sure what the risks could be in such an incentive, but geographic distribution is paramount to achieve real decentralisation.

One issue I can think of is more instability in the network (nodes opening up and down) as node hosts may chop and change location to chase higher incentive rewards.

I think the incentive would need to be very small, so a new host would benefit from hosting in a "dry geographic location" but an existing host would not be incentivised to move to that location.

@will-miningstore
Copy link

will-miningstore commented Oct 17, 2019

After engaging in conversation with Loki Developers @KeeJef and @jagerman, as well as reading the responses above and communicating with the Mining Store community, my opinion is left below.

First of all, I believe eliminating PoW was the right move for the following reasons.

  1. PoW was not necessary for the Loki project and therefore it was a waste of developer resources. We have seen Loki change algorithm several times since inception in order to remain ASIC/FPGA resistant, and the hours spent on maintaining an element of the project which I believe is unnecessary was wasteful. Furthermore, the PoW was heavily centralized as 80% of Hash rate was with a couple of mining pools.

  2. I agree with @jagerman that this issue is not about changing the emission curve it is about making a change that is best for the "technical consideration for the network".

  3. I think option 2 is the best decision moving forward as from my knowledge it does not affect the original emission curve, however, it only affects the incentive model. I believe that adjustment to incentive is positive both for network stability and for price appreciation for the following reasons.

a) A higher reward to service nodes will result in more users wanting to stake on the network and set up nodes. This will reduce the supply further and also increase the node count. Higher node count means a more secure network.
B) The reduction in circulating supply will cause increase price, and so to will the demand for Loki, as you need Loki to host nodes.

Option 1 might increase price because of the reduction in rewards, but this is manipulating the original emission curve. It also will not lead to the benefit of increased node count and therefore network security.

I am open to all opinions and I do not think my input is superior to anyone else, I simply have left my input for others to see.

Keep up the good work Loki team!

@smore12
Copy link

smore12 commented Oct 17, 2019

It seems that tail emmissions would once again be changed if option 1 is chosen.. Price of loki shoud not matter , is there a solution for tail curb emmision for option 1? Changing blocks from 26.6 current to 14.8 seems drastic.

for this reason or until something is brought up to curb the emmissions I vote option 2

@jagerman
Copy link
Member

In our case I think there is something to the idea that there is a greater propensity for miners to sell LOKI than stakers: I think that a significant majority of miners are pure profit seekers and dump right away. Sure, there are a few people mining to accumulate the coins, but I don't think they make up a huge share of PoW-derived rewards. People earning rewards from PoS, on the other hand, already have made a committment to the project, already know what it's about, and have already decided to hold some LOKI in it -- and so I think are much more likely that they will hold more of the reward than miners would. Or to be a little more economic-jargony, miners have a much greater marginal propensity to sell rewards than stakers.

Thus I think that if we take 1 LOKI from miners we would expect to see daily LOKI being sold drop by a little less that 1 (say, 0.95 LOKI), but if we turn around and give that 1 LOKI to SN operators/stakers we'd probably see a noticeably smaller increase (say 0.3 LOKI) in amount being sold daily -- and thus I think that even if we went with the full shift in rewards from PoW to PoS (i.e. option 2) we'd still see a reduction in supply (though of course not as large as under option 1) but also an increase in demand.

@Skelaton4
Copy link

I would vote for option 2, I am not in favor of changing the emissions curve again.

@GraftSpy
Copy link

GraftSpy commented Oct 17, 2019

A couple points that I have not seen raised:

Option 2 dilutes the premine faster than Option 1 (and at approximate the same rate of dilution that currently exists) while increasing Loki-flows to the LF; Option 1 would slow the rate of premine dilution while maintaining Loki-flows to the LF at current levels.

Option 2 would increase the reward for running a SN, which I believe is desirable -- particular if/because node operators will need to increase their VPS capabilities when Lokinet is live.

@Haafingar
Copy link

I have a lot to add to this discussion, I am nearly finished with my so far 12 page long diagnostic into this problem...bear with me!

@spacecatpixel
Copy link

I am biased towards option 2. Don't mess around with the social contract. This is the emission which was agreed to and the whole basis for your crypto economics paper. Risk is however economic incentive if LOKI price goes too low and it become not viable to run node. So my suggestion is option 2 with staking requirement increase.

Keep in mind that option 1 is the closer option to what was promised to SNs. Option 1 gives SNs the same amount of rewards so removing the mining emission completely results in the initial study still being relevant.

Option 2 in essence bumps the rewards the SN receive and option 1 keeps the rewards the SN receive at the same level as the economic study.

@spacecatpixel
Copy link

Every time I think about this problem it seems that both sides have their pro's and cons.

In my mind option 2 seems like the better solution due to:

  • Increase of SN count due to the larger circulating supply through time
  • Increase in SN count due to the larger reward.
  • Dilute premine at the promised rate.
  • Foundation reward doesn't change.

My only hesitation with option 2 is the emission rate. I've always thought that if we ever get the chance to reduce the inflation of the coin then we should. Basic supply and demand right.. Decrease inflation/emissions and thus the price increases. This raw statement is true however it neglects a bunch of other relevant information. I think inflation is irrelevant and daily emission schedule is more important and should be looked at.

Our current $USD daily emission is considerably low compared to other coins.

I'm leaning more to option 2 and here is why: Our current $USD daily emission is considerably low compared to other SN coins.

I've collected some data from masternodes.online, masternodes.pro and coingecko.

Note the data shown below is only a snapshot of these cryptocurrencies for the 29/10/19.

Coin AVG Block Time Blocks/day Block Reward Coin emission/day Price(USD) Usd emission/day
Loki 1m 59s 725 28 20300 0.3 $6,090.00
Dash 2m 37s 549 3.11 1707.39 73.49 $125,476.09
Zcoin 5m 288 50 14400 5.02 $72,288.00
NRG 49s 1,741 22.84 39764.44 2.4 $95,434.66
PIVX 58s 1,467 5 7335 0.25 $1,833.75

If option 2 is chosen (the mining reward is given to Service Nodes) then our daily USD emission is only $6,090 which means the price of Loki:

  • Could increase 20x before it has the same USD daily emission as Dash.
  • Could increase 11x before it has the same USD daily emission as Zcoin.
  • Could increase 15x before it has the same USD daily emission as NRG.

In my mind the USD emission/day gives us a loose understanding of the selling pressure a coin will experience. This coupled with the sentiment that Service Node's are less inclined to sell the Loki they receive, we do not have that much to worry when looking at our daily emission.

I've added PIVX to show that this finding may not be true as they have a lower USD emission/day than Loki however they are not increasing in price. I'm under the impression that Loki is more innovative than PIVX and sits within the ballpark of SN coins such as Dash and Zcoin.

TLDR: If our demand is growing quicker than these other coins and our USD emission/day is lower than theirs then I can see our price increasing quickly, regardless if we remove the mining reward or give it to the SN operators.

Therefor my vote goes towards Option 2.

@pailakapo
Copy link

I like the paper tries to keep a lot of the fundamental parameters constant as much as possible, as the economic model does have a large number of moving variables.

  • Rewards for Exit Nodes (Excluded from current paper only makes sense)

On the dynamic road map on the website, the lokinet mainnet transition is scheduled for EOY 2019, and exit node functionality sometime later.

Originally the back of the napkin proposal was for exit nodes to receive 2x the rewards of non-exit SNs.

If so, and even if the difference is not this drastic, the economic assumptions of this paper for SN ROI are changed as soon as exit nodes are given a reward over and above the regular SN reward, unless the emission schedule is also changed. And this could happen in the next 6-8 months.

@Lucifer1903
Copy link

@lewisphil that's a good point.

Am I correct in my understanding that it's not that exit nodes will receive double the reward form the current situation but that non-exit nodes are going to have their reward cut in half?

It would be good if the analysis took into consideration that relay nodes are going to receive half the reward in the future. The current economic analysis only works if we assume all service nodes will be exit nodes and receive the full reward, as opposed to relay nodes that will receive half the reward.

@pailakapo
Copy link

pailakapo commented Nov 14, 2019

I think the current paper deals with the block reward and SN/LF split. Once that is done, the next step would be splitting the SN to non-exit and exit ratio. Twice the current reward is not on the table, one of the options above is the reward.

So if they choose 20 loki / block reward, the exit node reward would be 2/3 of 19 (12.6 loki), and non-exit would receive 1/3 of 19 (6.3 loki).

It would be good if the analysis took into consideration that relay nodes are going to receive half the reward in the future. The current economic analysis only works if we assume all service nodes will be exit nodes and receive the full reward, as opposed to relay nodes that will receive half the reward.

And the ratio of non-exit nodes to exit-nodes is expected to be significant enough that assuming everyone is an exit node is not ideal.

@Lucifer1903
Copy link

@lewisphil I think analysis of exit node and non-exit node reward should be done at the same time as block reward analyst.

The analysis is assuming that service nodes will receive a certain amount of loki for a set amount of nodes on the network. If service nodes are going to receive different amounts when exit nodes come online then that makes the current block reward analysis pointlessly as the rewards will be different.
Only by considering the differences of exit node and non-exit node rewards into the current block reward analyst can we make an informed decision on what the block reward should be.

@pailakapo
Copy link

@lewisphil I think analysis of exit node and non-exit node reward should be done at the same time as block reward analyst.

The analysis is assuming that service nodes will receive a certain amount of loki for a set amount of nodes on the network. If service nodes are going to receive different amounts when exit nodes come online then that makes the current block reward analysis pointlessly as the rewards will be different.
Only by considering the differences of exit node and non-exit node rewards into the current block reward analyst can we make an informed decision on what the block reward should be.

I agree, which is why I brought it up. I was just saying what I thought the reasoning was.

@Lucifer1903
Copy link

Lucifer1903 commented Nov 14, 2019

@Haafingar I know you've spent a lot of time working on the analysis, maybe this is a lot to ask but would it be possible to do another one that takes into account the difference in rewards that exit and non-exit nodes will receive instead of basing the analysis on the current situation where there's only one type of service node?

@Haafingar
Copy link

I'm afraid it's not really possible to say much about exit nodes at this time. That's at least 6 months away for a start, but the bigger problem is that we just don't have any data on how willing node operators are going to be to run exit nodes. The amount of human work is going to be quite a bit greater than a standard node, as they'll have to be able to deal with requests from.

The issue with exit nodes is that the server providers can shut them down quite quickly, so it may not be advisable to run the entire service node on one machine to avoid getting completely de-registered if the machine running the exit gets forcibly terminated.

I really haven't done enough research into what kind of additional rewards operators would accept for the additional work and costs. The final design for Service Node exit provision hasnt been explored fully yet either, so in my option it is too early to factor in a potential design into this discussion.

I will however say that the original idea of exits receiving double the reward probably wont end up being implemented exactly like that. Assuming that we want a third of nodes to be exits, we'd have to know what reward they'd want in order to correctly carve out a reward for them. Off the top of my head, I'd stipulate that we'd likely want to add an additional stream off the 20 Loki per block we are considering now, but that's a later discussion.

@William-E-Coyote
Copy link

William-E-Coyote commented Nov 19, 2019

@Haafingar
You have stated that, "Assuming that we want a third of nodes to be exits..." and " the original idea of exits receiving double the reward..." Therefore one can assume that exit nodes will be receiving approximately 1/2 the emission of all SN.

So here is an idea: Wait until Loki Messenger is up and running as a functional product. Then market the rebrand of LM. Then implement check-pointing and simply turn off mining rewards. No reallocation or increasing of rewards to SN. With a reduced emissions the LOKI price will trend upward. Also at this time increase marketing for the new LM and make sure you state it is based on Loki Net technology. Also state that exit nodes will be active soon. Your marketing will coincide with increasing price of the coin, creating synergy therefore increasing adoption and awareness.

Then, when exit nodes are ready, give the mining rewards, 45% to exit nodes. Many new people will help create the exit node network, further increasing security and decentralizing the network.

@Lucifer1903
Copy link

What @William-E-Coyote says sounds like a good idea.
It's quite easy to to reduce the emissions but increasing emissions is difficult because of the social contract.
If it's clearly stated that mining emissions will temporarily stop and be allocate to exit-nods at a later date that would keep the social contract intact and is probably the most flexible option.

@pailakapo
Copy link

@Haafingar
You have stated that, "Assuming that we want a third of nodes to be exits..." and " the original idea of exits receiving double the reward..." Therefore one can assume that exit nodes will be receiving approximately 1/2 the emission of all SN.

So here is an idea: Wait until Loki Messenger is up and running as a functional product. Then market the rebrand of LM. Then implement check-pointing and simply turn off mining rewards. No reallocation or increasing of rewards to SN. With a reduced emissions the LOKI price will trend upward. Also at this time increase marketing for the new LM and make sure you state it is based on Loki Net technology. Also state that exit nodes will be active soon. Your marketing will coincide with increasing price of the coin, creating synergy therefore increasing adoption and awareness.

Then, when exit nodes are ready, give the mining rewards, 45% to exit nodes. Many new people will help create the exit node network, further increasing security and decentralizing the network.

I like this.

Even simpler, just assign the miner reward to exit nodes now. There are no valid exit nodes, so nothing now, and when they come online, the economic and social aspects are accounted for.

@Lucifer1903
Copy link

Even simpler, just assign the miner reward to exit nodes now. There are no valid exit nodes, so nothing now, and when they come online, the economic and social aspects are accounted for.

I vote for this idea.

@Haafingar
Copy link

Even simpler, just assign the miner reward to exit nodes now. There are no valid exit nodes, so nothing now, and when they come online, the economic and social aspects are accounted for.

I actually hadn't thought of this. This is worth consideration...

@KeeJef
Copy link
Collaborator Author

KeeJef commented Nov 20, 2019

@William-E-Coyote Service Node checkpointing is already enabled, the issue is we cant remove the mining reward before Pulse #9 is finished because we need someone to create blocks and order transactions, we could cut some of the mining reward now without it affecting anything too much. But if we removed the whole mining reward new blocks would not be created.

@Lucifer1903
Copy link

@KeeJef
He said now but I think he meant when pulse is implemented.

@William-E-Coyote
Copy link

William-E-Coyote commented Nov 20, 2019

Just trying to start a discussion for future emissions that include Exit Nodes (EN). I understand that there are many unknowns about how EN are going to effect the network, yet it is known that EN will make a dramatic change to the emissions schedule. Any accurate prediction needs to include the understanding that EN will cause considerable changes to the emissions schedule.

I like the idea of just shifting the mining rewards to EN. Simple and easy. Other ideas may be:

  1. Shift the mining rewards to a wallet which would be used to incentivize EN at a later time.
  2. Go with present plan B and state that you will shift the remainder of the mining rewards to EN.
  3. State that you are interested in starting conservatively and will shift the relationship of emissions to EN as necessary to support the network.
  4. All three of these work together.

I appreciate the open communication and professionalism of the team and community.

@pailakapo
Copy link

pailakapo commented Nov 20, 2019

@KeeJef
He said now but I think he meant when pulse is implemented.

Yes, by “now” I meant when the plan being discussed is executed.

@CryptoFirefly
Copy link

Reading the various comments on the thinking for Exit Nodes (ENs), does make sense to have a slightly longer term vision on the emission. If for the ENs we need to again increase the emission then we may want to re-think the 20 Loki proposal now and let the emission curve stay as is and never change. Rather think of more ways to burn loki for inflation concerns.

One of the thoughts after thinking and reading on getting the right incentives for the ENs is to do the following

  1. We do not know the right answer on what's right for ENs, so we admit that and park the item. But not exclude it completely from the assessment and economics.

  2. Keep emission at the rate we think we'll able to also reward the ENs and SNs to keep a healthy network and baseline that one time.

  3. If 20 Loki emission may be too low to split between SNs and ENs, (5% for LF), then we need to probably keep the final emission at 28 loki as is and not change the emission curve at all now.

What we could do is come up with a dynamic policy on re-allocation of the emission in %'s and we model and switch based on what's required.

While we wait for Mining to be dropped completely, we could start with giving a higher share to SNs now, maybe 75 SN: 5 LF : 20 Mining to begin with as role of miners would be dropped soon, so we can give that indication now probably along with launch of BLINK when SNs take on more work.

When POS is launched, we goto 95 SN : 5 LF (if the EN's are still a little while away on launch of POS).

Once ENs are launched, we arrive at the best model for SN : LF : ENs, (keeping LF at 5%).

EN's to earn more yes, but we need to model how many ENs are needed and model correctly.

E.g. if we want 10% of the nodes should be ENs, then we model to have ENs earn in range of twice to thrice of the SNs, but the split across would then look like 75% SNs : 5% LF : 20% ENs.

We could have 2 reward Queues, separate for SNs and ENs. Lesser ENs means, they earn more often than SNs.

  • So if we have 1000 SNs and 100 ENs, each day, the ENs would actually earn 3 times of the SNs in the above example, as 75% is divided among a 1000 SNs, while the 20% per block goes to a 100 ENs.

Per day earnings on a 28 Loki emission with 1000 SNs & 100 ENs would be
SN - (720 x 28 / 1000) x 75% = 15.12 Loki per day
EN - (720 x 28 / 100) x 20% = 40.32 Loki per day (i.e. x 2.66 of SNs)
LF - (720 x 28 / 1) x 5% = 1008 Loki per day

So big question is the risk reward ratio among SNs and ENs, and the % splits across should only be used to manage that keeping emission curve as-is.

@KeeJef
Copy link
Collaborator Author

KeeJef commented Nov 26, 2019

After reading and thinking a lot i think both options will have similar long term effects, but i want to make a few points which i think are true even though they might contradict each other

  • Assuming a fixed demand for Loki, transferring new rewarded Loki from miners to Service Nodes is likely to decrease Loki sold on the open markets in the short term (Based on the assumption that Service Nodes tend to either restake or hold Loki - or at least dispose of their Loki more slowly) This should naturally lead to an increase in Loki price on the open market in the short term.

  • Decrease in the overall new Loki produced (rather than a reallocation) would have the permanent effect of decreasing Loki supply, leading to reduced selling pressure. This is likely to make Loki more attractive to external parties looking to buy or stake coins.

  • An increase in the block reward given to Service Nodes is likely to encourage greater allocation of funds into the Service Node network. Once a 55-60% lockup ratio is reached, this increase in allocation is likely to diminish as the upfront cost of allocation becomes a barrier to capturing elevated returns.

  • Decreasing the Loki block reward more slowly dilutes the premine

  • Unless change is absolutely necessary sticking with a pre-agreed upon economic scheme is better than changing, which can lead to uncertainty about future economic guarantees.

I still have not decided what i think, although i'm leaning towards option 1 (Remove miners reward and do not reallocate to anyone, 14.73 total block reward)

@Lucifer1903
Copy link

@KeeJef are you leaning towards option 1 from the start this discussion (15 loki per block) or option 1 from Simon’s paper (28 loki per block)?

@pailakapo
Copy link

With an eye toward exit node rewards the miner reward going to service nodes short term makes the most sense.

Im not sure the 20 loki per block has enough room for exit node reward while leaving non exit nodes room for profit. Probably does, but its less certain.

@William-E-Coyote
Copy link

Overview:

This discussion goes back to the social contract. LOKI has stated its emission schedule. Changing the emissions schedule is changing the social contract. Although the social contract for LOKI is a bit more flexible than say BTC it still reduces confidence in the project whenever you make a change.

Exit Nodes are an important part of the project. It has never been clearly stated how EN will change the emissions schedule, therefore this has always been a time bomb in the social contract. I recognize that the previous idea was to give EN twice as much emission per node than a SN.

Pulse is a addition to the ecosystem that removes mining. When the project first began there was not any way to circumvent mining. Pulse allows mining to be circumvented and this increases security.

Because Pulse removes mining, this will be a clear change to the emissions schedule and the social contract.

Present:

Exit nodes are in development and scheduled for mid 2020. Pulse is in development and scheduled for mid 2020. Perhaps you could roll out both of these technologies together.

If you roll out Pulse before EN, you create a strong effect on the significance of emissions. Either you increase the amount for SN or remove the mining emissions and decrease the total emissions. Reduction in supply should create a higher price of coin. Either way, LOKI SN stakers win. Therefore this is a great problem to have if you are a SN staker or just a LOKI Holder.

If you Roll our EN before Pulse, you have a three way split and reduced emissions to present SN stakers. Then you still have to decide what you will do with emissions when you implement pulse. Not the best idea.

If you roll out both together, this is a much more complicated hard fork, but you can just shift the mining portion into the EN. You may also want to shift the percentage between EN and SN emissions. You can even just give 2x SN emissions to EN and see how it goes for a while before you further shift the percentages.

Conclusion:

Regardless what you choose to do, Pulse will improve the network. SN operators and stakers will make more money. Yes, the social contract will have to stretch a bit, but you cannot implement Pulse without stretching the social contract. Pretty much everyone wins except for the miners and they will just switch to another coin.

@Lucifer1903
Copy link

I agree with @William-E-Coyote, it's a good idea to implement pulse and exit nodes at the same time.

@William-E-Coyote
Copy link

Another Idea

After thinking about it for a while, maybe starting Exit Nodes (EN) before Pulse would work. Instead looking at the three way split as a problem, maybe it could be a solution.

You can slowly shift mining emissions to EN. If you know you are going to reduce mining emissions, why not do it gradually. Start with 20-50% of mining emissions and see how it goes. This gives you the ability to make further adjustments if necessary and implement Pulse when it is ready.

We really have no idea what is going to be necessary to support EN. More importantly, we really have no idea how users are going to use LOKI.net. If the project has poor adoption, then we have a problem. If the project grows beyond our wildest dreams, we have another problem. Regardless, we won't know until we get there. When we get there, we may have to shift the emissions to meet the demands of the users. And I really appreciate that the Team and Foundation opens up these discussions to the public before making a decision.

I still think that it really doesn't matter which way you decide to go on this situation. Everyone will be a winner.

Happy New Year!!

@dwarner5522
Copy link

I posted on Telegram but figured I would post here as well. My idea is to keep emissions the same but distribute the block reward as follows:

During the interim (before exit nodes are ready)
Distribute each block reward - 90% SN, 10% LF - This should increase the number of SN operators and also gives a 5% boost to the foundation.

Once Exit Nodes are ready the reward is as follows:
The reward split for a Standard SN would be: 60% SN, 40% LF.
The reward split for an Exit Node would be 90% SN, 10% LF.

So the LF gets a 5% increase regardless but makes out better if none of us make our SN's exit nodes. The investors make out better if we take greater risk and turn our SN's into Exit Nodes.

I know some people do not like the fact that the LF would be taking that high of a percentage, but I think that could give a needed boost to help marketing and overall development of the various platforms. The other option would be to just burn 30% of the reward when it's a Standard SN. This way the foundation always just gets 10%.

@William-E-Coyote
Copy link

Another idea.
Do we need Exit Nodes (EN)?
Presently we have Sessions messenger. Sessions is a SNApp that runs on the LOKI network and developed by the LOKI team. There is no reason that another developer will use the LOKI network to create their own private messenger app. Actually, the LOKI project is designed to promote this activity.

The strength of the LOKI.network will be to build an infrastructure of privacy SNApps. If LOKI is branded in such a way that it almost literally means privacy, it will become a trusted source for developers and users.

If we have EN, anyone could use any messenger app and increase their privacy. Furthermore, anyone could use any internet service with increased privacy. This would increase the traffic over the net, but not really help build the LOKI.network. It seems very plausible that an individual could use the LOKI.network for a BitTorrent client and claim it was the ultimate private BitTorrent client. And they would be correct. This BitTorrent client could use a significant amount of resources over the LOKI.net.

So for now, perhaps we should not offer Exit Nodes and we should build our own privacy network full of useful SNApps. Perhaps in the future, this could change, but for now, it is difficult to see the importance of Exit Nodes.

Maybe I am incorrect... totally understand. But just throwing another idea into the hat. I am actually pretty jazzed about Pulse and would prefer it to happen sooner than later.

@Lucifer1903
Copy link

I guess a good question would be how are exit nodes going to provide value to the network. It's easy to see the value it provides to users but what premium services is it going to provide to burn coins? Will these coin burning services be enough to offset the coin emissions rewarded to service nodes?
We currently have plans for coin burning service in respect to Session and the internal lokinet websites. However I have yet to see ideas regarding coin burning services for exits to the clear-net

@Lucifer1903
Copy link

That being said I think if premium services can be provided for exit related traffic then the freemium model is the best option to increase Lokinet's market share.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
ORC Loki Request for Comment
Projects
None yet
Development

No branches or pull requests