I currently run a node that hosts both the participation keys for my regular account and my folks escrow account. The stake in my regular account is currently 0 since the algo was moved to my escrow account, which is fine - the regular account simply does not end up being selected to vote or generate blocks.
Af the time I burn my galgo, the algo is returned to my regular acount. That move does not impact consensus because the algo is simply shiifted from one online account to another. Both before and after the redemption, both accounts remain online.
Iām not sure where the vulnerability is occuring with such a setup?
Thats not how folks finance works. Before you can redeem your galgo your folks finance escrow is emptied by folks. You do not move your algo from the escrow to your main account by burning your galgo, folks does that for you. This means that there will be a time after folks has withdrawn your algos but before you have burned where your online stake will be 0. This is true for all folks governors, which means that all algos online through folks will be offline for at least some time.
Because when you burn you burn against a central pool with all algos, and that pool cant exist unless folks moves all the algos from the escrow to that pool, which they do in just a couple of blocks.
Thanks. Would folks taking that central pool online themselves resolve the issue in your view?
Secondly, could the amount suddenly removed ever leave the network with less online than if they didnāt incentivize in the first place? If not, wouldnāt the network still remain more secure overall, even with a temporary drop?
As a representative of AF I assume your priority is to preserve the integrity and security of the protocol at all times, as number 1 priorityā¦ This proposal is clearly late notice outside of defi committee and has not been given sufficient time for other defi builders to prepare for (with exception of FF and possibly gard).
Given the nature of the galgo-algo redemption mechanism I have yet to hear a clear explanation about how the added risk from this predictable cyclical event can be properly mitigated.? With the profit motive of incentives, it is logical to assume existing non incentivized consensus participants will reorg to benefit from the defi 1m which as you admit will be clustered initially around FF at least for 1 or 2 cycles due to their headstart. For that reason alone it defeats its stated purpose and may increase consensus risks for the protocol at specific times above the status quo.
Please be patient and do this properly for future gov cycle supported by more defi options which donāt all share galgo-algo redemption timing implications.
On a personal Ievel I really appreciate what folks finance has done to bring attention to consensus and I think a reduced amount could be risked on defi specific rewards, maybe max 250k. But much preferred would be alternative distribution mechanisms across all participants in interim before the protocol level enhancements arrive. Please let @scholtz and others prove what they can do quickly and let others review as FOSS, under a Hackathon incentiveā¦
For anyone reading all these comments from the most active builders in the space like @d13co , @BunsanMuchi, @Swaggelwander , @scholtz (outside of FF/gard) and @lobo as a leading defi advocate. all are in agreement on the issues (late notice) and potential risks with this proposal.
Nb. @Adri , @trekianov - Collaboration is really great between AF and active teams, when it looks like collusionā¦ itās not so great, neither is knee jerk proposals. Please note I think this is only how it looks and Iām certain it has been well intentioned.
The only antidote to that on a truly decentralised platform is 100% transparency and proper due notice.
Can the AF committees (incl defi) publish all their meeting notes as matter of course? Simple audio recordings posted onto AF website within 48hours would prevent these situations and save lost times due to necessity of giving due notice on proposals so people are not wrong footedā¦ We all are feeling impatient but security is paramount. As it stands i do expect this to synchronize the loss of algo in online participation at predictable points in time. Until that is properly mitigated/resolved/explained it must be delayed or much less proposed to be put into the defi consensus particpation pot ( eg. <250k).
From what Iāve read here, my understanding is that the more successful the proposal is, the bigger the security drop. So even if this proposal works great, it creates an equally great problem because the stake will dump at a known time. Even in the optimistic case where we returned to the exact prior levels, itās not good to have drops like this. Moreover, what happens in 2024 Q1, Q2, Q3ā¦? Is this really a sustainable way to get people to participate or will we create a pump and dump in security?
I think consensus participation incentive structure is an integral part of the network protocol and should thus be handled by INC. We donāt have to try and tinker with these just because Folks added rewards and slightly more stake came online.
*** UPDATE 28 August ***
Thank you all for your feedback. Your deliberation has highlighted the need for further discussion on this topic. We will not proceed with measures 5, 6 and 7 this quarter.
I want to thank everyone who participated in the (constructive part of the) discussion and I would like to stress that the decision to not proceed in this quarter must be seen in a positive light: the Community has demonstrated enormous interest in addressing this issue and therefore itās necessary to have a more detailed discussion about the actual implementation and its consequences on the security of the network. At the same time, I feel the need to ask everyone to discuss, even fiercely, the ideas but abstain from transforming them into a personal issue: letās put aside our egos, we are here to build something together.
Really appreciate the decision to not rush it and instead trying to work on points that were brought up!
It is important to continue such a discussion now openly instead of it going back to a committee imo so we donāt end up with the same problem next period. This should be the general approach: get early feedback and try to work on the measures together.
And to comment on the last point: while my opinion was always backed by reasons (at least i hope i could make that clear) I know that some of my comments were directly directed towards FF. As long as we can a real discussion where we both acknowledge the points of the other party, balance pros and cons, I definitely donāt want to continue it this way. On top, part of this was due to the short time we had to make our points clear where I personally felt the need to address points more aggressively and I honestly just became pretty emotional because I just want the absolute best for Algorand.
itās vital to emphasize the preeminent significance of upholding Algorandās security. The ongoing fluctuations in market dynamics have inadvertently magnified the susceptibility of the blockchain to adversarial actions, thereby underscoring the imperative nature of its protection. The adverse aftermath of a successful attack not only endangers invested capital but also engenders the potential for widespread job dislocation.
As FYI, there is another idea/suggestion for participation node rewards here. To be clear, that shouldnāt necessarily block or impact this measure for GP8, but it is good to be considered for future decisions.
Idea:
The rewards (Say all or portion of fees + any extra Algos the foundation adds) go to a single participation node through a lottery (in parallel or combined with the committee selection) in each round, independent of its Algo stake size . Basically, receiving rewards works like an actual lottery and the only required qualification is running a participation node. In other words the rewards are given based on allocated hardware/network resources.
and this would lead for node runners to form pool coalitions so and distribute the rewards to coalition members
btw, how do you define the node runner? at the moment it is the account which is online not the nodeā¦
i think we really need to teach everyone that online account does not need to mean that you have to run the softwareā¦ Lets make 10B algos onlineā¦ for this we need to change this mindset
regarding the idependent thing of stake sizeā¦ i can split 1000 algo account to 1000 1 algo accountā¦ i can make it online in one nodeā¦ do you think i should have higher chance to win the lottery?
btw i agree that this lottery solution might be easier to implement and it might have also better legislative and accountings impacts
sounds like people will be incentivized to bring as many accounts online as possible which doesnt sound like a good idea imo. if you add a minimum stake to it, depending on the amount, it would just mean that whales can play this game
There is a cost to running a node and given that it is a lottery you would not start spending a lot of money running tens of nodes with the hope that the future payments will cover the costs.
