Reducing GLMR self bond and active set

Governance Democracy Proposals
,
1

Abstract

I suggest to reduce GLMR self bond to 100k and simultaneously decrease the active set to 72 seats

Details

Moonbeam self bond increased over the years from 10k→100k→2m GLMR. The main idea of this increase was to have more stability in the active set and avoid situation where collators goes in & out frequently

Once we reached that stability, we encountered another issue where we didn’t have enough collators to fill the set. Therefore we agreed (after the discussion attached below) to decrease the self bond back to 500k GLMR in order to attract more collators. Having said that, not many joined as expected since the self bond reducing wasn’t aggressive enough

Today, we’re in similar situation where the number of Moonbeam active set is 76, but only 69 are occupied. We can mitigate this issue in three different ways:

  1. Decrease only the active set to 69 seats
  2. Decrease only the self bond to 100k GLMR
  3. Do both - decrease the set to 72 and the self bond to 100k GLMR

Personally, I’m in favor of option 3 because it provides a good and reasonable solution to this case. We decrease the set, but still give opportunity to 3 collators to join (which will be filled instantly this time due to the low price)

Reducing the self bond only without reducing the set (option 2) will definitely cause instability among those 7 collators who will try to catch their spot. Hence, it’s better to minimize it to 3 seats. On the other hand, reducing only the set to 69 (option 1) will hurt decentralization

Therefore, as suggested above, I believe the golden road is doing both, and on the way we also match between the set size (72) of both Moonbeam & Moonriver

Will be happy to hear your thoughts!

Best,

Rafael

Links

2 Likes
2

Dearest Rafael as you well remember from Iceberg Nodes we were already in favor of this direction some months ago… so for sure yes, count on our support about this topic, totally. Numbers are there and are quite clear: we need such kind of countermeasure👍

3

Hi Michele,

Yeh, I totally remember you were in favor of reducing the set, and now we have a better opportunity to do it together with reducing the self bond

Thanks for your support!

Rafael

1 Like
4

Seems reasonable. Let me pass this on to the dev team

3 Likes
5

So the reduction in the self bond will be added to RT4100 which will come to Moonbeam in January.

For the reduced active set, this can be done directly through governance. I will post a proposal later in January after the TC elections complete and the Term 3 delegates are put in place.

3 Likes
6

Thanks Aaron for the update. I believe it’s a good step to stabilize Moonbeam collator set by doing those actions.

2 Likes
7

Hi @legendnodes Appreciate you revisiting this suggestion. Respectfully I disagree and do not believe the bond should be set to 100k GLMR. Currently we can all agree the market lately has impacted everyone and every block chain so no surprise growth appears stalled for new collators. Setting the bond too low grants a better entry vehicle for folks whom will not take validating seriously and professionally. I’m not against revisiting the bond amount but I think currently it is too soon since the last reduction to assess the impact due to current market conditions.

8

Hi @blackk_magiik , yeh, I remember you were against it even the last time I suggested it, and even though you expressed your concern, I honestly don’t see it as an issue. Let me explain why and reminding everyone the process and the logic behind my suggestion:

1. We discussed my previous suggestion 9 months ago (and the reduction took place ~7 months ago), which is a decent period of time to have a conclusion if the move was good or not

2. Back then I suggested we reduce it to 100k GLMR (from 2m), and eventually we all compromised to lower it to 500k (which is still high imo but better than 2m)

3. The main purpose was to have more collators in the set because many left and we had empty seats to fill

4. However, as expected (to me at least), NONE of new collators have joined even after the reduction to 500k, which confirms it’s still too high (allow me to not include the collators I added out of courtesy to share my part to the ecosystem)

5. Moreover, few more collators even dropped since then, where 4 of them were with 2m self bond that might have stayed if the self bond was 10k or 100k GLMR - we’ll never know

6. The current stage is that we have a set size of 76 collators and only 68 are occupied. This doesn’t concern you? How do you think we should fill the gap? Or do you believe we should keep it like this and hopefully it’ll get better somehow in the future?

7. As I explained in the past - I don’t think we should keep the self bond too high and make the current active set like a private & centralized group. I do want to encourage new collators to join

8. Therefore, I suggested decreasing both the set and the self bond, in order to hopefully get more collators but still limiting the size of the set and align it with Moonriver. If we just reduce the set, then let’s reduce it already to 68 because no one will join if the self bond won’t be lowered accordingly (especially if the market will recover, because it’ll be much more expensive in $ value to join). And If more active collators will drop then we’ll continue to reduce the set to 60 & 50, etc? At which point do you think we should change our approach and reduce the self bond to make it easier for new collators to join?

9. Personally, I don’t see any reason why we should have different self bond per collator. I think it should be the same for everyone like the lower bar of 10k GLMR. I’m sure that if someone proposed that each collator (including the active one) should have a 500k self bond - you would immediately vote against it, so why do you support it for newcomers? because they are new? Why do you first assume that there are higher chances that bad operators will join rather than good one? I actually think exactly the opposite.

10. Having said that, let’s assume that you’re right and non professional collators will join - we still have the tools to remove and chill them out in 12 hr. Just have a look at the inactive collator list and you’ll find some of them such as MoonThailnd from the early days and stellaswap from the recent weeks. Both (and many others) didn’t perform well and got chilled - simple as that

11. This is why I’m not concerned at all if some poor operator will join, but I’m more concerned if collators will keep dropping and no one will join because we insist on a 500k GLMR self bond

12. Beyond that, I don’t see any reason that the current collators should be worried about the newcomers because there are available seats. The newcomers (if any) will compete against each other, not the old one which has a much higher stake.

13. If after all these moves still no one will join then we should consider reducing the set accordingly imo (whatever the occupied seats will be by then), but first we need to make it really easier for newcomes to join.

14. Don’t forget that even current collators can add few more (up to 4) collators if they wish, and tbh - I hope it’ll will happen because anyone who survived by now and still operating is a very good candidate to me

Final word - I totally respect your decision and don’t underestimate it even though you’re against it. I admit that I honestly don’t fully understand it but again - totally accept & respect it. I have no doubt you want the best of the network but we both have different approaches on this topic, which is always good to have different thinking where we can share thoughts

Best,

Rafael

1 Like
9

Hi @legendnodes thanks for posting and always putting a lot of consideration in how we can make Moonbeam successful. You present some good options and I agree that something needs to be done. On your options, do nothing is also an option, which because of the fact that collators who are not producing blocks can be side-lined by a call, a collator that has given up operations can’t negatively impact chain I think is still viable But, I also feel that if a collator has given up, their spot shouldn’t be “saved” by bond and the fact some delegators aren’t paying attention, so I don’t think ‘do nothing’ is a good option.

I think of the ones you presented, I would be in favor of a modification of 3, where the set is reduced to the amount of block producing collators (less than 69, currently), for the following reasons.

Due to current market conditions, I think stability for stakers is the most important consideration. Many of the nodes that have gone out now still have delegations where the staker is unaware, and those stakers will get an unwelcome surprise when they look back at their positions. There’s no doubt in my mind that of the collators that have stopped producing, some have done it because it isn’t profitable enough for them to continue operating their node, and I expect more collators will drop out impacting their delegators, too.

Because of this, I think the priority should be on ensuring maximum rewards for both current stakers and collators, and to do this, decreasing the size of the set to the amount of collators which are currently producing blocks (let’s define it as those that have produced blocks in the past month) will help accomplish this. To me, the bond doesn’t really matter at this point and can be decreases to something like 50k or even 100k as you suggested. But even now, with the rewards of collators only focused on active block producers due to sidelining the non-producing collators, we see that it appears Blockdaemon is may be exiting the position on one of their collators possibly to cease operating the collator.

In fact, I see this modified change to option 3 as helping to keep the set decentralized. In continuing to maximize rewards for existing, established, and proven collators by shrinking the set to those that are participating, eventually there will be a point where the rewards are sufficient that there will be no more drop outs. However, if the bond is too low, and we reserve spots for “decentralization” without a clear indication of what a number for decentralization is, it actually has the opposite effect, enabling a single entity to swoop in and take slots as more drop out - akin to the Amazon effect where their costs are low enough that they can wait out other entities to fold. Currently of the block producing set there appears to be a good amount of varied members so we should try to keep it this way as much as we can.

A case could be made that by keeping the spots unfilled it could put upward pressure on GLMR because new entrants would purchase GLMR, but this would only be temporary. Also, one could say it could create a fight situation putting upward pressure on GLMR, but it is known that this isn’t good for stakers from previous experiences do the the constant changes in the active set. If a new entrant with high enough conviction wants to join the set, the total GLMR to snipe the lowest collator that is producing blocks is now low enough that it would only take a modest amount to do it.

So to me, action should be taken, and it should be taken as I presented to ensure stability, over everything else.

2 Likes
10

Further to my previous message, now it seems P2P is exiting and I think my point about lowering the set size is worth an another look.

2 Likes
11

Yes, that was the plan from the begining - reducing the self bond and the set to 100k & 72, respectively

I think we just need to wait first for a few weeks and see whether reducing the self bond bringing new collators act accordingly. Meaning, if no one joins (or very low new participants), then we can reduce it even lower than 72, but let’s wait first and see how it evolves with the new minimum self bond

2 Likes