Revisiting a Sector Duration Multiplier (FIP-0056)

[vkalghatgi]

Author: CryptoEconLab

Problem Motivation

Earlier this week, @Stefaan-V started a discussion on current challenges and economic headwinds the Filecoin network faces. Some key issues include:

• A challenging investment environment given current global macroeconomic dynamics coupled with recent volatility in crypto markets
• The lack of incentives and ability for providing longer-term storage on the network, given that long-term network commitments are crucial to demonstrate the network’s permanence with regards to valuable data storage
• Storage Provider sustainability given the current environment, including uncertainty surrounding liquidity and future network rewards

The following proposal is by no means an exhaustive solution to the above list of issues/challenges facing the Filecoin Network, but could help better align the economic incentives of the network with longer-term storage commitments, attract capital interest to the ecosystem, and improve the network’s stability while providing Storage Provider’s with further optionality and reward potential.

The proposed mechanism draws on elements of FIP-0036, and the discussions and analyses that followed. They are our take on a maximally agreeable subset, selected to bias towards achieving community consensus and minimizing delays, while still being net positive for the network and individual stakeholders.

Proposed Mechanism

• A linear Sector Duration Multiplier (SDM) will be introduced for all sectors, including Committed Capacity (CC) sectors and sectors containing storage deals. The SDM will have a slope of 1x.
  • A longer sector will have a higher Quality Adjusted Power than a shorter sector, all things equal. The multiplier increases the amount of FIL expected to be won per sector per unit time based on the duration the sector is committed for. This introduces the idea that longer term sectors represent a long-term investment and storage commitment to the Filecoin ecosystem, and therefore should be rewarded proportionally with greater block reward shares.
  • The multiplier enables Storage Providers to express, and receive rewards for a bullish outlook and long-term resource commitment to the network. Furthermore, operational burden and risk to Storage Providers increases with duration, and it is fair that they’re commensurately rewarded for this.
  • The motivations for a linear slope of 1x are that any nonlinear function results in more complexity without clear gains, and SDM’s with slope less than 1x do not sufficiently compensate for the risk incurred from longer commitments and do not provide sufficient benefits to the network.
  • The Duration Multiplier is multiplicative on and independent of the existing Quality Multiplier incentive (Filecoin Plus incentive). Filecoin Plus already offers a 10x multiplier for proving storage of data from verified clients.
• Increase the minimum and maximum sector duration commitments of 6 months and 1.5 years respectively
  • The motivation to increase these is decreased sector turnover and improved network stability. Given that the network expects to grow, having sectors that expire every 6 months means that the network potentially needs to find new sealing throughput to compensate for the loss of power from expiration. This hinders the network as it continues to scale. A related discussion that proposes to extend the maximum sector duration can be found here

We present some scenarios for implementation of the SDM policy, followed by a discussion of the tradeoffs relevant to these parameter decisions.

	Minimum Sector Duration	Maximum Sector & Deal Duration	SDM Eligibility	SDM Function (SDM slope is 1x for all options)
Scenario 1	1 yr	3.5 yrs	SDM only applies to sectors onboarded after the network upgrade that implements this FIP	Linear slope of 1 starting at 1.0 years
Scenario 2	1 yr	3.5 yrs	All sectors (SDM is applied to existing sectors upon upgrade and extension)	Linear slope of 1 starting at 1.5 years
Scenario 3	1 yr	5 yrs	SDM only applies to sectors onboarded after the network upgrade that implements this FIP	Linear slope of 1 starting at 1.0 years
Scenario 4	1 yr	5 yrs	All sectors (SDM is applied to existing sectors upon upgrade and extension)	Linear slope of 1 starting at 1.5 years

Parameter Spaces and Tradeoffs

Two main parameters for the community to consider are (1), if the maximum sector commitment should be increased to 3.5 or 5 years, and (2), whether the SDM policy only applies to sectors onboarded after the policy's implementation or if the multiplier also applies to already existing sectors at upgrade and extension. These options each have respective benefits and risks which are discussed below.

1. Increase the maximum sector commitment from 3.5 years or to 5 years.

• 5 year sectors are preferable from a longer-term commitment perspective, and also enable storage services that offer longer-term data storage
• A maximum sector duration of 5 years introduces better stability and locking dynamics to the network
• 3.5 years may be safer as it is more incremental, and could ease concerns regarding higher collateral needed for committing sectors for longer-term durations

2. Whether the SDM policy should only apply to sectors onboarded after the policy’s implementation or if the multiplier applies for existing sectors upon upgrade and extension

• Applying the policy at extension enables Storage Providers to express longer-term commitments without incurring the computational cost from resealing sectors. Allowing the multiplier at extension also enables some sectors to function as a sponge to soak up liquidity when extended, reducing circulating supply and initial pledges for other sectors.
• Allowing the policy at extension may lead to fairness concerns between sectors with FIL+ deals and sectors without due to access to the multiplier. While there are arguments that applying the policy upon extension may be net positive for sectors containing FIL+ deals, the determination of whether this is net positive should be part of the community discussion.
• The multiplier only applying to newly onboarded sectors guarantees equal access to the SDM regardless of sector type. Furthermore, the duration multiplier can start at 1-year instead of 1.5 years since there is no comparative disadvantage to sectors already committed for greater than 1-year. If the policy were to apply at extension, quality multipliers from longer-term durations would only apply for sectors committed for greater than 540 days.
• Enabling resnapping is already a priority for Filecoin devs. However, enabling the policy at both sector upgrade and extension increases importance of implementing resnapping to support parity in access to the multiplier between sector types.

Next Steps

Following community discussion on these policy changes, CryptoEconLab will look to introduce a FIP ideally for inclusion in the upcoming Network v18 upgrade with guidance from the governance team at the Filecoin Foundation as well as Filecoin core-devs.

[dmc314]

I think this is a good point worth highlighting:

The proposed mechanism draws on elements of FIP-0036, and the discussions and analyses that followed. They are our take on a maximally agreeable subset, selected to bias towards achieving community consensus and minimizing delays, while still being net positive for the network and individual stakeholders.

It's also worth highlighting that this does not include any proposed changes to the collateral target locking parameter.

[momack2]

I would like to argue for scenario 4 (5 year sector max lifetime, applied to all sectors) - because I think its parameter set is most aligned with long term network value and capability. To break down those two tradeoffs:

1. 5 year sector max lifetime: A max sector commitment of 5 years best supports the needs of a range of storage clients who want long term commitments to data preservation (a number of archival use cases see 5 years as the minimum). A 5 year maximum also implicitly supports the 3.5 year option that some storage providers and clients might compromise on, without holding back our most bullish SPs and clients from being rewarded for an even longer term commitment.
2. applied to all sectors: More strongly than the max lifetime - I STRONGLY think it's in the best interest of the network to enable SDM to apply to all sectors. We currently have 15 EiB of capacity sealed into the network, and 385 PiB of user data. If these existing sectors aren't allowed to benefit from the SDM multiplier, it creates a strong incentive for SPs to terminate or expire the existing sectors and either exit the network (bad, releases collateral, hurts all other network participants) or reseal their sector (bad, wastes energy and sealing time that could be dedicated to network growth). This increases the chances that SPs collectively can't keep capacity onboarding above network growth expectations set in the baseline, which would decrease overall block rewards for all SPs. ⚠️

Re 2 - I think there's a potential compromise that mitigates concerns for deal-making SPs.
The problem: I understand that deal-making SPs are concerned that if they extend the max sector lifetime of a sector that is already filled with deals, they won't be able to harness that sector for future deals (to get the full-length fil+ multiplier) once the initial deals expire (which given past max deal lifetimes had a max length of 1.5 years that the protocol can't unilaterally change).

A specific example: I have an existing 1.5 year sector with a 1.5 year FIL+ deal in it. I currently get a max quality multiplier of 10x. After SDM, I can extend that sector to 5 years, but I still only have a 1.5 year FIL+ deal, so my power multiplier would only be 10x5 = 50x for the first 1.5 years of this sector, and then it would drop down to just 5x for the remaining 3.5 years.

The solution: There is significant work in flight towards "Re-snapping" sectors with existing deals, to replace an expired deal with new data and apply the new deal multiplier for the remaining length of the 5-year sector. Core Devs could prioritize this FIP to land it in ~ex a network upgrade shortly after SDM takes effect, in order to bring long-term usage of extended sectors to deal-making SPs that need/want to use their data storage resources fully.

Updated example: I have an existing 1.5 year sector with a 1.5 year FIL+ deal in it. I currently get a max quality multiplier of 10x. After SDM, I can extend that sector to 5 years, but I still only have a 1.5 year FIL+ deal, so my power multiplier would be 10x5 = 50x for the first 1.5 years of this sector. After 1.5 years, I either make a deal with the same client to renew their FIL+ deal (same data stays, additional deal lifetime of 3.5 years is negotiated), or I can choose a new FIL+ deal to replace the old expired deal, and swap it out. Either way, my multiplier stays at 10x5=50x for the full 5 year sector lifetime, maximizing my benefits / advantage.

Curious to hear from folks if they agree that Scenario 4 seems most aligned with the network's long-term value/capability, and whether this mitigation assuages deal-making SP concerns with choosing that option. 🙏

[anorth]

Unfortunately a few details make the analysis regarding trade-off (2) not quite right.

We have to break down pre-FIP-0045 deals (which account for almost all current deals) from post-FIP-0045 deals.

• Pre-FIP-0045: The sector cannot be re-snapped, due to complexities of the QA-power calculation that FIP-0045 resolved. When extending the sector, the existing multiplier is spread out over the new, longer commitment so that it is constant for the life of the sector. The multiplier doesn't stop when the deal expires. By example a 1.5y sector with 10x multiplier can be extended to a 5y sector with 3.7x multiplier; or the SP can wait for the 1.5 years to elapse, then extend the sector with a 3.5x multiplier for the remaining 3.5 years (the first option earns more total reward, but more delayed in time). But they can't put a new deal in the sector.
• Post-FIP-0045: The sector cannot be extended beyond the deal term until/unless the client takes action to extend the maximum term of the associated datacap claim (the default max term set at deal creation time is deal duration + 90 days). If the client does do this, the SP can extend to 5y and enjoy 50x multiplier for the duration 🎉 . If they do not, the SP cannot extend and their multiplier is just 10x. When the deal minimum term is reached, they can extend for the remaining 3.5y and take a new deal at that point (again, requiring the client to increase the maximum term if we're still dealing with the built-in storage market's limitations at that time), with a multiplier of 20x for that remaining period.

After SDM, I can extend that sector to 5 years, but I still only have a 1.5 year FIL+ deal, so my power multiplier would be 10x5 = 50x for the first 1.5 years of this sector

This conclusion is off. A multiplier on a FIL+ deal can only scale with the duration for which that deal is committed. So a 1.5y FIL+ deal is only 10x (given the multiplier function of scenario 4). Otherwise the SP could get a 50x multiplier for some period despite only hosting verified data for much less than 5y. The only way to get 50x is to commit to a deal for a 5 year term.

For pre-FIP-0045 sectors/deals, the "fairness" objection isn't changed by re-snapping, because these sectors can't be. It may be possible to migrate these sectors into a post-FIP-0045 regime, but when we considered this initially we found significant compromises. When re-snapping is implemented, we could have another look at that – perhaps there's a scenario where an SP could drop a deal which has completed out of their sector, even though some rewards have not yet been earned, because the potential to re-snap a new deal is more valuable. This potential value wasn't a consideration earlier.

[anorth]

I would like to open discussion about the possible incentives to terminate sectors if the SDM did not apply to extensions of existing sectors.

If these existing sectors aren't allowed to benefit from the SDM multiplier, it creates a strong incentive for SPs to terminate or expire the existing sectors and either exit the network (bad, releases collateral, hurts all other network participants) or reseal their sector (bad, wastes energy and sealing time that could be dedicated to network growth).

I am not convinced this is as bad as this claim (which I know is held more widely).

Regarding incentive to terminate existing sectors, I would like to see it quantified. I think it would have to be based on some kind of scarcity argument, e.g. the SP doesn't have the storage capacity or pledge funds to make a new SDM commitment, so they have to terminate an existing non-SDM sector to free it up.

• Constrained by storage capacity: By assumption, the reason they would do this is to commit a new SDM commitment for the same capacity, the net effect on raw byte power is zero, but with higher pledge lockup.
• Constrained by pledge: There's nothing to be gained, since short-term return on pledge is not increased by a multiplier.

If an SP were not previously incentivised to terminate and exit, I don't see how the introduction of an SDM option would increase it significantly. The only way could be through wide adoption of the SDM reducing their share of rewards, but wide adoption is a success case and means much more storage and pledge were committed by others!

Regarding re-sealing, I agree that the energy cost is probably a net social loss. However, I don't think sealing time that could be dedicated to network growth is a valid concern. Almost by assumptions behind this proposal, SPs are not limited by sealing capacity. We've observe much higher network sealing throughput, growth is low now for other reasons. That sealing hardware is probably not otherwise being used toward network growth under the conditions that motivate this proposal.

This could be really good. If, in fact, we can determine that incentives to terminate existing sectors are actually weak or zero, that gives us a wider range of options. There would still be reasons to prefer applying to existing sectors, but we can evaluate them for the potential up-side. It would be a shame to have chosen this only in response to a worry that wouldn't have materialised.

[f8-ptrk]

there is always an incentive to terminate sectors: getting the pledge out for liquidation. the higher the pledge the higher the incentive in the end. adding a duration multiplier without adjusting the termination fees in correlation to that multiplier will increase the risk of people terminating sectors to "cash out" the pledge. as rewards are expected to go down over time we even have a bigger problem here as the gap between the 90 days termination fee and the higher earlier rewards due to the multiplier gets more severe than with non multiplied sectors

[tmellan]

@f8-ptrk The termination fees are proportional to duration up to the fee cap (via SDM scaling power scaling rewards). So the current relative value of penalties to incentives should be good for a few months before there's much change compared to today. Then it can become an issue as you point out. So it seems like we have that as a grace period to design as careful a choice of new termination fee structure as possible. Does that make sense?

[tmellan]

Initially I thought of incentives to terminate as a primary part of the justification for an ‘all sectors’ approach, however I agree this doesn't seem majorly convincing (and so it was never pursued by me at least).

In my mind one of the main reasons now for ‘all sectors’ is to minimise barriers to maintaining network capacity — there's been a fairly significant drop here, down around -1.4EiB so far. By putting up barriers to all sectors extending, some implications are:

• If you’re concerned about shifting the security base away from hardware that’s one thing.
• Lower locking than otherwise could be achieved is another, which affects economic security of the network, but also brings a host of other issues.
• And another point is SDM access shouldn’t be viewed in zero sum terms, as it would lower the pledge for verified deals being onboarded in the future.

Having said that, is there any clear issue with applying the SDM only to new sectors now, then if we’re feeling more confident, bringing (or not) SDM to ‘all sectors’ in a subsequent upgrade?

[anorth]

I don't think there would be any additional challenge in bringing "all sectors" in a subsequent upgrade – it would just shift any work we need to do for that back in time. As applied to already-committed sectors, it would reduce the maximum multiplier they could attract, since they'd be a bit older.

[jennijuju]

to ‘all sectors’ in a subsequent upgrade?

Other than what anorth said, I think a couple of other things were raised, that might be worth addressing before we can consider "apply to all sectors":

• network reaction to the CE shock
• fairness between "how much work needs to be done to get more incentive" -- fast resnap?
• Fil+ program was flawed and we dont want to incentivize existing bad behaviors -- FIL+ move to l2 and have its own punishment mechanism?

[f8-ptrk]

there should be an option to opt out of the multiplier when doing long term sector pledging.

[dmc314]

Having a variable SDM would introduce some complexity in the code base and in analytics/tracking. We'd also want to consider if the potential benefits outweigh the simplicity of the original proposal.

[tmellan]

To maximise the largest proportion of commitments being able to opt for the longest durations this seems sensible. To my mind the question is should it be immediately, or later, as an improvement after more consideration?

Trying to think through considerations:

• Additional implementation work as Dave mentions (though, may be a small factor).
• The incentives become more complex and harder to reason about. Although it seems fine there may be things we don’t understand yet about how this would play out having not given this structure a huge amount of attention. (This is the main reason to delay adding this feature imo).
• And on delay, if it turns out that not having the optionality is a limitation and there is demand it could easily be enabled in a subsequent upgrade.
• Another view: If the justification is to bring ‘affordability’ to duration, then the optionality seems not absolutely necessary. An SP can selectively choose not to make use of the deal multiplier, as needed. But I’m not convinced circumventing the limitations of simple fixed SDM like this is quite the right logic or in the spirit of what this proposal is trying to achieve.

[colorman0404]

赞成引入退出机制，退出后销毁前期乘数奖励代币，否则无法退出

[momack2]

Another flag on added complexity from a user perspective over a SP-set multiplier:

• clients reasonably want as much of a guarantee as possible that storage providers will abide by their commitment to their storage deal.
• Therefore, it would make sense for clients to have a preference for 5 year sectors with a maximum multiplier over 5 year sectors without the multiplier (and therefore less collateral, less termination fee, etc).
• However, if it's up to SPs how much collateral they add to a deal, then clients need to be able to see the "amount staked" for a deal when deciding which storage provider to work with (as part of the deal parameters?)?
• Seems like that would add a lot of complexity to the user experience of choosing qualifying SPs for their deals, and also to the number of groups that would need to adjust deal-making flow to enable that decision

[gitjduck]

With regard to the user complexity perspective, a few counterpoints:

• Nothing about this is simple - with or without SDM, with or without variable SDM
• To bring some degree of simplicity for clients would be to eliminate the obfuscation of second-order incentives like Fil+ and SDM, and have clients simply... pay FIL... for storage and/or retrieval based on the purpose and value of the data
• SDM manipulates network incentives, it doesn't materially change client/user guarantees. Nothing about SDM stops an SP from shutting down retrieval after ingesting data, or charging an arbitrary amount for the retrieval. Until the protocol is more robust, long-term client/SP relationships will require person-to-person interaction to achieve practical guarantees.
• This also means "users" will very rarely be network clients; users will instead broker their activities through an intermediary that abstracts network client complexity away (again, precisely because it is already very complex, regardless of what happens with SDM)

If Filecoin L1 presumes to anchor its incentive structure to longest possible storage duration, it risks permanently binding itself to the subset of use cases it could address in its infancy - rather than the variety of use cases protocol improvements might enable over time. 5 years is a long time in tech, particularly if you realize after 6 months you used the wrong approach.

Philosophical aside: Power multipliers have a Midas' touch quality. They solve problems in the short term, but the effects are hard to anticipate - and it is impossible to go back. Once SDM is solidified, should the network face new cryptoeconomic challenges, it won't be tenable to reduce power for SPs that have collateralized PiBs of 5-year Fil+ deals... so the way forward will be to multiply power again based on some new, even more justifiable (?) criteria.

We agree added complexity is undesirable - let's start pursuing simplicity by putting down the magic multiplier stick.

[anorth]

Thanks @vkalghatgi and your team for restarting this discussion and presenting a wider space of tradeoffs for us to explore together. I'm glad these ideas are back in active discussion and look forward to us landing something into the network soon. I hope and expect that the restricted scope and presentation here are much more widely acceptable.

I think the general thrust here is positive for the network and almost all stakeholders, especially in the long term. I have some concerns about further shifting the security base of the network towards tokens rather than hardware, and that we might be over-paying in terms of power/reward to secure longer commitments, but they could be seen as letting the perfect be the enemy of the good. Everyone's idea of perfect here will differ, and it's valuable for us to quickly settle on something good so we can learn and iterate if warranted. These proposals are close enough for me.

I am fairly confident in supporting this proposal in scenarios 1 or 3, i.e. new sectors only.

Scenarios 2 and 4 (applying to extensions of existing sectors too) present tradeoffs that are more balanced. In addition to the fairness issue (which can't be resolved simply by fast-tracking re-snapping), these scenarios bring somewhat more risk of volatile dynamic behaviour, the potential for a race etc. Some prior proposals suggested a ramp on the multiplier to mitigate this, but I am hesitant to introduce complexity here. I would really love to just see some "what if" modelling demonstrating stability if uptake is higher than expected.

I think there is potential to design around the fairness issue with a combination of re-snapping and an opt-in migration for pre-FIP-0045 sectors that has re-snapping in mind, but that would take this squarely out of the realm of a quick and easy implementation, and sooner + simpler is probably better. Doing that work would have opportunity cost pulling experienced Filecoin engineers from other long-term utility-oriented projects. I am not sure the fairness issue should necessarily block applying this to existing sectors, but the tradeoff definitely has a downside.

[momack2]

I think that scenarios 1 and 3 are actually harmful to network dynamics, since it increases incentive for expiring and terminating sectors from the network, and I think that is a much worse outcome than a temporary feeling of unfairness which is actually much much smaller than the net benefit that improved cryptoeconomic impacts has to all SPs.

[geoff-vball]

In scenarios 1/3, SPs would expire sectors in order to roll that pledge into new (presumably) 5-year sectors. In scenarios 2/4, SPs are still incentivized to expire some sectors in order to extend their others. In either case, the net onboarding/offboarding should be roughly equal, shouldn't it? Or is there a reason that wouldn't be the case?

[momack2]

why would SPs be incentivized to expire any of their sectors in scenarios 2/4? If they can raise the funds for the collateral, don't they have an equal incentive to extend ALL their existing sectors?

[f8-ptrk]

pledge is not equal on all sectors. old sectors tend to be way more "expensive" from the pledge perspective. and since pledge is measured on the original pledge amount new sectors are most likely way more FIL efficient than old ones

why pay x5 pledge at 0.2 fil/sector if you could get x5 on 0.02 FIL/sector now? (example numbers)

pledge calculations need to be adjusted, sure. different topic thou

independent from that, dependent on deal mechanics (sector quality, resnapping, deal extensions) old deal sectors are most likely doomed for sdm multiplier extensions

[geoff-vball]

@f8-ptrk I don't think that is correct. Pledge is recalculated on extension based on current network dynamics. You don't get a refund though, if the new pledge calculation is lower than the old one.

@momack2 If sealing a 5-year sector costs 1 FIL, I already have X FIL pledged in my sectors, and I can raise Y additional FIL, then at the end of the day I'm going to end up with about X + Y sectors. I don't see how this is different between any of the given scenarios. In the case extending is allowed, these sectors are going to be less favorable because extended sectors won't be eligible for a 5x multiplier - if my sector has been on the network for a year, the best I can get is a 4x multiplier.

[Irmagogo]

Is it possible to set SPs into different classes with storage capacity as a variable condition?
Firstly, find the average QAP of SPs across the network and use that QAP as a benchmark. Divide a rank for each certain number of QAP, and increase the SDM slope by the corresponding percentage for each increase in rank until 1.
For example, the average QAP of each SP in the network is 60P. Assume that the rank increases every 20P. The slope of SDM of SPs with more than 60P QAP = 1. The slope of SDM of SPs with less than 20P = 0.7; the slope of SDM of SPs with 20-40P = 0.8; the slope of SDM of SPs with 40-60P = 0.9. (The numbers are just examples and do not represent anything of value)
In that case, if a completely new SP wants the easiest way to join the network to store the user's data, and the SP chooses the minimum time of 1 year, that's 100.7=7x. If the SP chooses 3.5y, that's 103.5*0.7=24.5x.
Given the current secondary market and lending situation, entering the network with fewer pledges may attract more SPs.
The benefit is that small SPs can enter the network with lower trial and error costs, while big SPs can choose how to grow on their terms.

[geoff-vball]

I don't think limiting smaller SPs to lower multipliers would be good for diversification or growth of the network. I do like the principle of @f8-ptrk's suggestion of letting SPs pledge "up to" a certain multiplier, giving them choice on how much stake to input and a corresponding QAP gained. I think the only objection to this idea would be technical complexity (which may or may not be worth it).

[stuberman]

It is inspiring to see that the ratio of real data stored on network is growing relative to the CC stored. How does this proposal prevent that relative growth from being reversed? A reversal would be harmful to the value of Filecoin to the world, especially in light of the FTX debacle where policy makers question the value of crypto systems.

I, too, question the concept of "incentive to terminate existing sectors" when trying to justify additional incentive to maintain or grow CC sectors. I believe the opposite is true, with additional incentives for real data, CC Filecoin miners will actually be incentivized to snap real data into existing CC sectors. Storing real data that benefits humanity is the motivation behind Filecoin, not simply mining tokens.

[DecentCr8]

I have strong skepticism about whether "RealData" should be given more block rewards than "CC", and I am also skeptical that "RealData" is worth more than "CC" at layer 1. Here is the reasons:

• CC sectors are Storage Providers' assets invested into the network, which shows the external recognition of Filecoin Value, if you consider Filecoin itself as an island economy. CC storage are quite valuable, just as powers to Bitcoin.
• CC is the fairest mechanism in the network, that is to say, it is completely decentralized, determined by algorithms, and guaranteed by cryptography; It is the cornerstone of Filecoin blockchain security;
• On the contrary, "Real" Data is artificially identified, each copy is identified by a few people, it is centralized, and it is even difficult to verify. The more rewarding to "real" data power, the more threat to the blockchain security;
• Among all the current large amount of "Real" Data stored in Filecoin, vast majority is just to obtain block rewards and does not generate any actual data value. This may corrupt the Filecoin network since the power is governed by notaries (homans), rather than a decentralized network or algorithms;
• The current evaluation of the value of "Real" Data is indiscriminate, that is to say, the incentive is completely based on 10x computing power, even if the Verified Client sends completely garbage data, which is unfair in itself. This assessment should be left to the market.

A better solution is to move Fil+ totally to Layer 2 after FVM is ready. At that time, we could have Fil+, Fil++, or Fil* etc, each has its own token to incentivize, and compete each other. Each "real data" tokenn will be evaluated by the market automatically. This will be no security impact at layer 1.

This is actually is a repeat of my comment in #442

[Fatman13]

A better solution is to move Fil+ totally to Layer 2

+1

[jhookersyd]

Clients want 5-year storage deals, we need to facilitate what they want, not what we think...
Australia has 43 universities with LOTS of research data, 5 years aligns with their current data storage contracts.
Also, five years aligns with our hardware warranties.

[f8-ptrk]

then allow 5 year sectors and 5 year deals but do not force a multiplier on it. as soon as we do 5/5 there won't be any other deals for quite some time and everyone who won't have the money to stake even more, half a decade in advance will exit the network

the math behind this argument is elementary:

Someone has 20 FIL. if someone needs to stake 10 FIL to store data for X years he will store data for X years. If someone now needs 20 FIL to store the same data for X years then he will only store half if he hasn't got 40 FIL to begin with.

that will only lead to things we do not want:

• miners without unlimited cash will leave as the price for long term storage doubles
• miners will store less data as they have to take on a greater financial risk to do so

if we make the multiplier optional we might be able to combat the second, if not the first effect of the multiplier.

i will not again bring the termination risk up - it is there and will be higher the smaller the miner is. also simple reward math. this should really really be thought out more before pushing a duration multiplier to the network.

[nai2283]

I would choose option 3, regarding if the DM should be subject for SP to choose, I don't think we need to worry too much about it, since SP can still distribute their sectors among regular/FIL+ to control the overall cost. And from another perspective, if the token price raise 5 times, would've less storage been committed to the Filecoin?

[f8-ptrk]

since SP can still distribute their sectors among regular/FIL+ to control the overall cost

in the current deal market situation i doubt an SP can actually do that. The deals that are availabel on the "free market" - evergreen/spade are to be named here as an example - will most likely instantly move to 5 year, FIL+ only.

long term i do not dispute that SPs could choose balanced strategies with FIL+, durations etc. but i doubt a lot of miners would make it into that future if we have a period where 5 year, FIL+ is all you can get and you have to pledge double to X more than is needed

[ekil1100]

Scenario 3 will be great for flexibility. Also, I agree with f8-ptrk, there should be an option to opt-out.

[Destore2023]

We'd like to choose Scenario 3.
We need to optimize the economic model in the shortest term, so SIMPLE is the best.

[DcentHosting]

Scenario 3 with an option to opt-out of the multiplier makes most sense to me. But as F8 already stated, I don't fully see the reason for having the multiplier so high. It makes it very difficult for everyone to deal with. Personally I would do a x2 as a multiplier on 5 years. At maximum.

One thing we do love, the option for 5 years. This should be implemented asap.

[sjcdltx]

In general terms the concept of a sliding multiplier for length of duration sounds reasonable and I agree that scenario 3 is best. However, in alignment with @jhookersyd we should be considering this from the perspective of the "customer" not the SP.

Flexibility is key: If the goal is ultimately to enable real customers (public data or enterprise) to store data on Filecoin, retrieve it, and over time do much more with it (move from cold to warm to perhaps one day hot storage), then what we should be providing is flexibility of length of contract able to be adjusted to the needs of the client and not fixed centrally. With that, incentivising the client to sign up for longer term deals is in everyones benefit (including getting a better multiplier). In the enterprise world, we would sell contracts based on 1, 2, 3, 5, 7, 10 years.

We should be incentivising the positive actions for network growth, as per @stuberman - we absolutely should be incentivising the storage of real customer data (public and enterprise) over all else.

The elephant in the room alongside all this remains availability of FIL (and the fact the when it is available for borrowing, it comes with a 20% interest rate that almost kills the breakeven). This is for another thread/FIP but am sure this can be solved for elegantly...

[thatch45]

I would go for scenario 3, primarily because I think this has the least potential impact on destabilization and centralization of the network.
I hope I am wrong, but I suspect this will have the potential to increase centralization overall, scenario 3 makes the over centralization and destabilization risk lower. Since it will require the new model to be sealed into, rather than snapped into. The burden of compute being applied will stretch out the required timelines to take full advantage of the new multiplier.
Any time a large scale economic change allows for too rapid modification of the underlying economic structure it introduces stability issues.
Scenario 3 may incentivize CC sectors to terminate and re-seal, but it requires it to happen at the rate at which those sectors can reseal - which will be congruent roughly to the sealing to storage ratio of the miners - this means that it can't all happen at once, like scenario 4.
Scenario 3 also mandates that new deals be sealed, old deals could be re-sealed and recreated, but this processing overhead will slow it all down - which is very good.

3.5 years is not long enough, allowing existing sectors to upgrade without resealing is too dangerous - scenario 3 is the only option.

[jhookersyd]

So...

• Scenario 3 feels like the consensus here = Why? - KISS
• Opt IN/OUT for the FIL staked = SPs have a choice
• Updating staked FIL/duration on the fly = Great for all
• Terminating old sectors for new = Reallocation of capital is sensible. Investments aren't static.
• New sectors only = Fair for all
• 5 years = Requested by clients

Next stop 2 dollar FIL. Let's get this done people!

Many enterprise SPs will disappear if we don't get this done asap! (Holon is fine FYI)

Agree with @DecentCr8, FIL+ should be moved to the FVM at some point. This will make it easier to program flexibility into future models. "Genomics/Healthcare FIL+" feels like the natural first step for a dedicated incentivisation program. If we nailed this one vertical well, it would lead to exponential growth for the network.

[Aaron01230]

It is normal for the client to have a storage requirement of more than 1.5 years. In my opinion, PL should provide a more convenient renewal plan for this, instead of increasing the output multiplier based on the length of time. In addition, from the perspective of SP, in the absence of pledged FILs, the longest period of 5 years corresponds to at least 10 times the number of pledged FILs, and no funder will agree to their loan needs. So I refuse this proposal absolutely.

[PHABRIC77]

@Aaron01230 this is a good point. The network should aim to be as frictionless as possible given its broader goals.

My 2 cents on the above and the broader discussion: -

1. Network value is most important. Generating client fees is the only way to increase network value sustainably.
2. Any arguments about supporting the FIL price (for example through locking up pledge) without increasing the network value are fundamentally flawed.
3. Whilst I accept and acknowledge the concerns expressed above about SDM, we could write an equally damning verdict on CDM. Supporting one over the other ignores the key issues and is counter-productive.
4. The problem statement should be as follows:- How do we collectively ensure a sustainable increase in network value through setting appropriate incentivization for contributions (of varying nature) to the network?
5. This encapsulates many things - making the network more frictionless, introducing enterprise standards, building functionality, providing consensus etc etc. But The CDM vs SDM debate has always been about relative incentivization. Let's address this head on and be balanced and objective and mindful of the desire to increase network value.
6. Once we understand an appropriate incentivization model and agree its aims we can start to explore appropriate mechanisms of incentivization. Combining duration and a multiplier (leverage) is not the right approach as both things increase risk exposure. Leverage is bad at times (remember 2008) and duration risk is bad at times (SVB?). At worst the introduction of these two aspects should be carefully considered and only implemented with appropriate risk mitigation already in place (like a lender of last resort, a fluid and functional lending and staking market etc.). In short, these changes should be made when the FIL economy is functioning and not before. It is also always a question of degree and scale. 10x feels dangerous, 20x feels reckless. 5 years feels dangerous, 10 years feels reckless etc.
7. Why have other incentivization improvement mechanisms not been explored. For example, the ability to haircut the required pledge to incentivize certain standards or behaviors would serve to reduce the requirement for leverage. For example, if you half the pledge, you double the returns for a given level of risk. I recently asked the question why the pledge needs to be maintained at its current level there was no clear answer offered. Other possible mechanisms include favorable FIL lending terms, adjusting the baseline minting schedule etc etc etc

Debate is welcome, but it needs to develop to effective and rapid action. My understanding is that this debate has been too long, and without appropriate response and this is the problem that should be addressed urgently. Otherwise, how can the network hope to adapt to the ever-changing environment in the future (and thus help ensure sustainable network growth)??

[Irmagogo]

As a member of Singapore SPWG, I agree SDM should not go ahead, CDM has to be discussed.

"Widened belts are hard to shrink back to their original shape" The same applies to Filecoin Network.

[herrehesse]

Dear Filecoin Community Members,

Following in-depth discussions within the Global Storage Provider Working Group (SPWG), DCENT has determined its position against the implementation of the Sector Duration Multiplier (SDM) as proposed in FIP-56. We believe that this stance not only represents our view but also accurately reflects the majority of Storage Providers' perspectives.

The SPWG's core values encompass increasing network value, promoting permissionless access, and ensuring profitability for Storage Providers. We contend that FIP-56, which proposes the introduction of SDM, directly contradicts these principles and does not serve the best interests of the network. Instead, we believe that the Capped Duration Multiplier (CDM) is a more appropriate proposal to support the network's long-term growth and sustainability. This stance aligns more closely with our core values and has recently gained significant support from the active Filecoin and Storage Provider community.

Sincerely,

The DCENT Team

[flyworker]

FilSwan did a community research and discussion inside webchat group (378 user) in Asia. the majority opinion is against this proposal.

[Porterstowell]

Reminder: Last Call for comments regarding FIP56 is EOD March 24th. Please make sure you add your opinion to the discussion before then. After Last Call, the next step is the Community Governance call scheduled for March 27th 3-4pm ET. Register here: https://fil-org.zoom.us/meeting/register/tZMkf-2qpjIqEt3nkeKh_7f7_F6wPm76zbTw

[dkkapur]

Thanks again for the continued engagement and productive discussion on this FIP0056 proposal. As it moves to final call this week, on behalf of the Filecoin Plus governance team, @galen-mcandrew and I are sharing the following statement. We hope to address the challenges raised by the community and share more details on the actions that we are taking.

In light of all the current conversations and pros & cons to this FIP, the Fil+ governance team is confident that the community will continue to grow and thrive and therefore is not taking an explicit stance for or against this proposal. Nonetheless, we hear and appreciate a lot of the feedback and requests that have been voiced as result of the FIP discussion. We are working on a broader Fil+ community-wide statement which includes position for more specific points (i.e., what % of the community supports position X on issue A), and we expect to post another update by end of week from Galen on what the community aligns on as a broader group. Please reach out to Galen or me directly if you have any questions.

This FIP0056/SDM discussion, as well as that in the parallel CDM proposal, have continued to highlight the importance of Fil+ in enabling Filecoin to achieve its goal - building a data store for humanity’s information. Over the last year, we have seen a tremendous amount of growth in the Fil+ pipelines, enabling storage of 99%+ of all active deals in the network today representing a growth of about 50x in total data stored on the network. Our main focus during this time was to scale up the program - evolving the LDN process to its current v3.1 implementation which has drastically reduced time to DataCap, increasing the number of active notaries in the system, and supporting a wider range of use cases through E-Fil+ and other pilots. This has also shown us that ultimately, there are several participants in the program today that are abusing the system and looking at ways to maximize their earnings to the detriment of others doing the hard work of “correctly” onboarding data - winning business, preparing data correctly, storing with reputable SPs, and ensuring the data can be retrieved by the data owner/client/users. In the last few months of 2022, we made an explicit pivot towards working towards minimizing abuse in the system. This resulted in releasing the CID checker, working with individual notaries and data owners to improve dispute handling, and empowering the T&T WG to scale up. We see the mission of this program as “The Fil+ program supports use-cases for data onboarding and utilization of Filecoin’s decentralized storage network through community-governed trust mechanisms.” Based on this, we plan to continue down the path of prioritizing supporting quality data onboarding to the Filecoin network. This means that independently of how this proposal progresses, the Fil+ team is committed to the following:

• Working towards 95% of DataCap being used by trustworthy storage clients, which means reducing potential abuse of DataCap to ideally 5% or less. This is initially defined as cases where no actual data is being stored, data is being over-replicated, or applications are falsified - in essence, “false positives” for “trustworthy clients” or sybils attempt to profit disproportionately from QAP block rewards.
• Maximizing use case coverage for types of DataCap of applicants → better designing for the needs of data owners and notaries for onboarding open data vs. private data (and other slices). For all types of applications we have seen today, there should be a path for applying for DataCap or an explicit descoping from support.
• Pursuing decentralization, minimizing bias, and building more transparency. Committing to investigate proposals from the community (i.e., trustless notary) and collaborate on how we’re going to continue evolving the program together with other stakeholders.

In line with the above, here are a few current priorities that we’d like to highlight:

• (recently shipped v1) notary leaderboard - https://notaries.datacapstats.io/: the goal here is to help highlight notary activity and increase attention to notaries who are making the effort to support quality onboarding on Filecoin. This is the first step in the direction to establishing notary reputation and better tracking/rewarding appropriate engagement for this key stakeholder in the program.
• Incentive alignment and introducing skin-in-the-game for relevant actors in the DataCap pipeline. Experiments are being designed right now, e.g., Deciding the appropriate staking method  notary-governance#846, Notary collateral staking and incentives with FVM notary-governance#817.
• Evolving the DataCap allocation process design - another big step is definitely warranted (similar to the transition into LDN v3 last year). Lots of folks have proposals, we’re taking an active stab at evaluating and sharing outcomes, tracked here: Fil+ design exploration notary-governance#841. This also needs to account for supporting different type of applications, as mentioned above, and explicitly supports private datasets / enterprise use cases in a safer way by scaling up E-Fil+ and related pilots/processes.
• Improve the dispute management and resolution process. This starts with empowering the T&T WG to drive open disputes to resolution and ensuring the community is on the same page about how to productively flag potential abuse. Lots of ongoing topics on this front, if you’re paging in fresh, start here.
• Improving data collection and transparency in the system. This specifically includes setting better expectations for clients with regards to KYC/KYB (see more), improving the application process for both clients and notaries, as well as continuing to push more data publicly about the program (e.g., improvements to DataCap tracking, transparent logs for tooling).
• Improved consequences for identified DataCap abuse. This includes leveraging the tools we already have (i.e., CID checker) and are building towards, to react faster to potential abuse in the system and hold clients, SPs, and notaries accountable accordingly. This also includes tracking reputation of actors across time (clients, notaries, SPs, etc), as well as ensuring we have tools to leverage whats available to us in the protocol, i.e. removing DataCap from clients and exploring other QAP modifications in the future.

There are several other things in flight as well, please check out our roadmap! Restating that we are committed to ongoing evolution and improvements to the Fil+ program and processes as the state of the network continues to change. Our top priority is and remains to support onboarding from trustworthy clients. We’d love to hear your feedback and suggestions on the above in relevant GitHub discussions, an upcoming Fil+ governance call, or in Slack.

[15012700225]

Fil+ is the root of datacap transactions, the main culprit of network disorder, and I hope he can disappear. The person who proposed the scheme should apologize.

[Stefaan-V]

As @dkkapur mentioned above, the teams are committed to increasing the users and data quality of the network by implementing new tooling and automated processes.

In the last couple of months, there have been questions about our overall strategies as an ecosystem to drive and accommodate the growth of the network. As we are in the second phase of our master plan, we continue to prioritize our efforts on demonstrating utility and onboarding data to the network. Here are some of the other major initiatives, but not all, that are in progress as of March 2023. We will continue to update these as more details roll out.

1. Building tooling to support data onramps

Web2 apps and data owners manage the largest datasets today. To increase the adoption of the Filecoin network, we need to help bridge those Web2 applications and make it easier to store and retrieve data from our decentralized storage network.

To accelerate adoption of the network, we need to reduce the onboarding friction customers face when storing and retrieving data onto the network. Since mainnet went live, we have seen plenty of successful onramps built and deployed. However, the proprietary implementation makes it unavailable to the rest of the community that doesn't have the capital to hire engineering resources.

As our network is a decentralized storage platform, it requires the rails to store, track and retrieve data onto the platform. To be successful, we need to build a series of tools that can form the foundation to build the decentralized onboarding services and make those scalable, publicly available, and vetted by the community. In this way, we aim to create better customer experiences and help build standards on how data gets onboarded, indexed, and retrieved.

The Outercore engineering team, a division of Protocol Labs, has been building onboarding tools to enable onramps and SPs and made them available publicly here. We continue to collaborate with the ecosystem to improve these and vet those out further. Check out this landing page to learn more about these upcoming tools.

2. Driving more top-of-the-funnel demand

In the last year, the network has made a ton of progress in getting Web2 datasets on the Filecoin network. Today, Filecoin hosts datasets of some prominent Web2 businesses and many more copies of public datasets originally stored on web2 clouds.

Our main Web2 user interests today come from early adopters in research, academia, media, and entertainment. Though we have a long way to go to make Filecoin a household name in the Web2 space, we are seeing increased organic interest from Web2 decision-makers coming to our ecosystem. We need to build on these successes and continue to create awareness and drive demand for our Storage Provider ecosystem by educating the market on the benefits of Filecoin through paid campaigns. We will start these campaigns next month.

3. Reducing the cost of sealing for Storage Providers

One of our key initiatives has been and continues to be focused on reducing the cost of processing the data before storing the cryptographic hashes on-chain. Two immediate improvements are being released:

3.1 Reduce Sealing CapEx by up to 80% with new software and hardware configurations.

These new recommendations and software enhancements are built for existing sealing deployments and can achieve up to 80% lower cost versus most current deployments. This improvement results from collaborating with key industry leaders participating in the Decentralized Storage Alliance (dsalliance.io). These enhancements will be released and made available publicly in the next quarter. Keep up to date with the latest release timelines here.

3.2 Move to an OPEX model and leverage on-demand sealing services.

We are introducing sealing-as-a-service software that allows for a new ecosystem of Sealing Providers to join our ecosystem. These could be existing providers with excess compute capacity or new providers (e.g., Aligned) with optimized hardware architectures to offer sealing services to the Storage Provider community. Storage providers needing extra sealing will be able to consume these services on demand instead of using CAPEX to purchase compute resources. Keep up to date with the latest release timelines here.

4. Support and enable new DeFi staking programs to support SP growth

With Storage Providers increasingly storing more data onto the network and needing access to FIL collateral for pledging, we must ensure FIL liquidity moves easily through the network. In the last few months, we have worked closely with entrepreneurs to build liquid staking solutions on our FVM stack. We are committed to continuing to help and support these new offerings to develop and scale their stack on FVM successfully. We still have a lot of work to do, but expect that this new DeFi economy will provide access to millions of FIL liquidity to our provider ecosystem in the next few months.

Check out the blog here by Ashwanth and monitor all ongoing staking solutions listed here.

5. NEW Upcoming Economic FIPs

During this FIP discussion we received requests to look at solutions to alleviate the collateral requirements and prepare for uncertain dynamics in the gas economy. So the teams started a few new FIP proposals specifically to address these:

5.1 _Reducing Upfront Collateral (Project Network Shortfall)

• Problem

The storage capacity has been dropping in the last year as storage providers were redirecting their pledge to FIL+ Datacap or other regulatory requirements.
There is a tremendous value in increasing the allocated hardware capacity to the network via Committed Capacity and Fil+ alike.
To support these collateral challenges the team is exploring an economic proposal to entice more hardware commitment into the network by lowering the upfront collateral pledge barrier.

• Proposal

The Network Shortfall proposal would allow providers onboarding new storage capacity (or extending, snapping) at a fraction of the initial pledge requirement – a pledge shortfall. This shortfall is accounted for at the miner actor rather than the sector level.
This new implementation should kickstart and support a growth cycle in data onboarding and hardware capacity; while supporting a less inflationary supply trajectory without necessarily depending on large external capital inflow.
A more concrete proposal is being worked out as we speak. Please comment and participate on that proposal here.

5.2 Optimizing Gas impact through a Multi-dimensional Gas Lane

• Problem

With the launch of the FVM and upcoming IPC applications, there is uncertainty around the relative supply and demand of block space and different types of messages. BatchBalancer was a previous attempt to address a similar issue. Exploring a more general solution that works with varying levels of supply and demand is desirable.

• Proposal

Read up on the details from AX's blog here on how new gas lanes will help the network against a gas sprawl and help maximize the network revenue by optimizing the blockspace allocated for FVM messages.
Next week, we will publish a FIP discussion introducing separate gas lanes for SP messages and general FVM messages as a tool to hedge FVM-related gas risks.

[jhookersyd]

@dkkapur @Stefaan-V All the SPs that are voicing their concern against FIP56 don't want human based governance in a blockchain system.

Bitcoin has endured time due to its clear protocol-based incentives. CDM gives the Filecoin community this long lasting clarity to build a real business, selling real data, to real clients.

It is clear to see that PL just wants to push this FIP through which has NEAR ZERO community support.

Mark has made some great points on how FIL+ should change from where it is today. = filecoin-project/notary-governance#841 (comment)

TLDR

You can’t have ‘trustworthy client’ or ‘datacap allocation commensurate with earned trust’ at the same time as Decentralization. Great Notaries will be bias, bad notaries will be malicious. Adding more Notaries to FIL+ just creates a larger attack vector, again I think we are seeing this play out in real time.

Filecoin+ should be centralized and bias because it inherently will be, the program should be treated somewhat like and R&D grant whereby ANYONE can apply for datacap but the approval is based on the decision on a smaller notary group that represents stakeholders regionally. The focus of FIL+ should be on driving REAL adoption and Development and thus value to the network. Datacaps should not automatically be 10x but much more nuanced, a function of the value being driven by the data being onboarded. The ‘budget’ of such a program should be reviewed annually and scaled up and down based on the community’s view of whether it is successful (driving value to Filecoin) or not.

[dkkapur]

@jhookersyd i think there is a distinction between:

• folks that believe that Fil+ needs to evolve (including but not limited to ideas raised in that discussion) to support quality onboarding and minimize abuse
• folks that believe that Fil+ shouldn't exist regardless of how good it gets in minimizing abuse / false positives -> either due to the impact it has on SP business design or the extended argument where it puts the network in down a more subjective + PoS (vs PoST/PoW) path

LMK if you think the above is not a fair distinction, because I see people in both camps today. I'm more indexed to the former, mostly because it is more immediately actionable and there are tactical/strategic shifts we can push together for in the community. I think the concern raised by the latter is a fair one, but that is a broader question we should be aligned on at the network level. Filecoin's "master plan" still puts in the "onboard lots of data" phase, and Fil+ serves as a way to reduce friction for that, but I see where that group is coming from and I think its worth flagging explicitly.

Thanks to you and Mark for participating in that convo!

[dkkapur]

@f8-ptrk @SBudo @dcasem ☝🏽 would appreciate your thoughts on this too.

[f8-ptrk]

we have a problem as soon as we dive into "quality", "useful", "real" or similar categories regarding data. and in the end the FIL+ program itself should not care and by FIP design actually doesn't.

we have a problem with scale of trust. we should and most likely must limit the FIL+ program drastically. No idea why LDN exist (well, because some clients got unlimited trust, namely the PL platforms) we never should give more than 32 GiB (or so) of datacap to a new client and then scale slowly per client.

we have a a scale problem overall. at some point FIL+ becomes a problem consensus wise. we most likely need to limit FIL+ to a max % of overall raw power. Like no more than 2% FIL+ of newly onboarded raw power

we have a problem with how humans capacity scales. notaries cannot scale effort properly. other problems exist with notaries that needs to be solved too, sure.

Now there are discussions that certain SPs and PL are looking at applying 50PiB at once - and the answer needs to be no. wrong direction completely

[dcasem]

We should keep refining FIL+, even though I'm not very optimistic about its success unless we make some significant changes. However, it's crucial that we find a path to make CC sectors viable -- they help secure the network and allow for different business models outside of FIL+.

By further disincentivizing non-FIL+ data, we're going against the principles of a permissionless and decentralized blockchain. I'm concerned that passing and implementing this FIP, despite significant objections from the community, could seriously harm the project's future. Once we get past this issue, I would advocate re-evaluating how governance works, especially how FIPs are passed.

[SBudo]

I believe that FIL+ had and has its place within the network, but it drastically needs to change as you mentioned above.
There is a mid-way between the two options you have above, where FIL+ can be used for specific use-cases that helps developing the network by increasing its usefulness.

The biggest problem the network is facing right now is the power gap between FIL+ and unverified deals/CC.

The long-term success of Filecoin should remove the reliance on block rewards and shift to SP selling storage directly to customers. While FIL+ has the power it has compared to unverified deals/CCs, this will never be achieved as it does not economically make sense for anyone. FIL+ should be considered a short-term mechanism to get us to that ultimate goal while we develop all the necessary tools to become more competitive and attract customers.
Instead FIL+ has turned into the cornerstone of SP's survival.

The chain consensus is also a major concern. Once the total raw power of sectors containing FIL+ deals get to (today) 10%+ (or 5%+ with SDM) of the total raw power of the network, the majority/concensus will be achieved by FIL+ power. But FIL+ power is currently assigned subjectively by a set of Notaries. This in essence means that the Notaries will hold the consensus power on the chain (an off-chain process!), and they can arbitrarily assign or revoke more or less power to any clients/SPs.

[f8-ptrk]

@Stefaan-V @dkkapur

while your input is valuable - this discussion is in the end not the right place. FIP56 will be implemented regardless of developments in FIL+ governance or off chain tooling available in/to the community.

if there are concrete FIP'able steps resulting from the actions your teams take, please start new FIP discussions to give the community the chance to provide input on/for these proposals.

FIP56 is in the eyes of a lot of people a threat to the core of the network, its mechanics, its future and ideals - its impact going way beyond what off chain tooling and governance could potentially fix we need to be very careful to not justify implementing something with the destructive power of FIP56 on the premise of what might or might not be happening off chain.

[dkkapur]

ACK, fair enough. I still stand by sharing thoughts on this discussion topic since this is where a lot of eyeballs are coming and folks that are building opinions should see the full picture. I don't think its reasonable to state that someone should be for/against this FIP based on where Fil+ (specifically off-chain topics, like abuse in the current system) is at today, but then choose to ignore even its current plans. So either scope both in and get the picture, or scope both out.

FWIW I think the call-out to level up the discussion to be focused on network incentive design and philosophy / ideals is a good one. Just noting that that's not all people are scoping their current threads and opinion-forming to.

[15012700225]

Fil+increases sector computing power by a factor of 10 out of thin air and increases the multiple by extending the commitment time. In both cases, it is obviously more equitable to increase the computing power multiple through the commitment time。

[Fatman13]

Hello, all, I am the coordinator of CN SPWG group and we just held a meeting in light of Porter and Jennifer’s calling for FIP governace engagement. Here are some highlights of what we have discussed… (Note that things may get lost in translation)

TLDR

As there are still arguments on if SDM and CDM is exclusive to one and the other. The “vote” was collected on two front.

SDM

• In favor of SDM: 1
• Against SDM: 2
• Neutral（保留意见）: 9 ✅

With respect to CDM

• SDM + CDM: 1
• SDM: 1
• CDM: 10 ✅
• Neither: 1
• Neutral（保留意见）: 1

Discussion

• One SP shared that SDM and CDM is mutually exclusive due to their inherent nature. It would be hard to imagine that SDM being passed for May upgrade and CDM come around later to amend SDM. What would happen to the already x20 sector?
• One SP in favor of SDM shared that 20x is somewhat acceptable as x10 and x20 wouldn’t make too much of different in current network.
• One SP shared that SDM may very well do the opposite of convincing SPs/investros alike to commit to the network. Changing minimum duration from 180 days to what SDM is suggesting now would expose significant more risks to business that might be willing to join but now further deterred.
• One SP shared that storage clients would always want to store longer deals is a myth. Nobody it worked with has suggested that.
• One SP shared that the network should allow flexibility for SPs to join and quit even if they are opportunists. The more people are willing to join the network the better.
• One SP shared that the reason people stopped extending their CCs is that doing CC is no longer profitable. SDM would further drive this trend and weaken the security consensus of the network. Again SDM exposes more risks for business driving them away from the network instead of having those would-be SPs joining.
• One SP shared that the liquidity that SP sacrificed for SDM would put business at great risks.
• One SP shared the concerns of hardware life span.
• One SP shared that SDM essentially diluted the production cost of FIL which in turn may very well further decrease the fiat price of FIL as the pegging between hardware and filecoin value become weakened.
• One SP shared that concerns that SDM would drive more decrease in raw byte power.
• One SP shared that some core developers advocates the network has to do something to the cryptonomics before May upgrade shouldn’t be the reason to push for SDM. Old saying of 不变应万变, especially with regard to tokenomics.

Off-topic

• One SP had discussion on “verified data” of f+ always getting misinterpreted in Chinese as “真实数据”(real/valuable data). As of now there are ~600Pib verified data stored on Filecoin. If we are to make an analogy of Filecoin loosely as a company assuming that Filecoin as a Web2 company have control of 600P data, the company’s value would be through the roof. Whereas in reality, verified data adds practically zero value to the network rendering f+ program in a very awkward position.
• One SP shared that even without SDM being introduced a filecoin SP is already becoming more and more risky as the block return kept going down and gas costs now make up 20 to 30% of the block returns.
• One SP shared that with fevm now coming online. Many of the L1 complexity could now be experimented to be moved to L2 making a cleaner L1. To educate investors, btc/eth miners or would-be SPs on the protocol has already deterred a large number people.

[Destore2023]

Great Job， Fatman！

[kaitlin-beegle]

Thanks for sharing, @Fatman13

[jennijuju]

Thank you for a productive and insightful call - super focused & informative feedback of SDM for core devs IMHO. @Fatman13 great summary

[ChrisWardJones]

Great summary @Fatman13!!

[danlessa]

Hey'all. BlockScience has worked to perform a first-principles-based review and compare the FIP-0056 (i.e., SDM) and CDM proposals. It has been instigated and funded by the Filecoin Foundation due to the controversy generated by recent proposals and the consequent need for an independent technical group to read through the arguments, check them, and evaluate their implications. The full article can be found at Medium. Our Executive Summary, Diagnosis and Recommendations on the whole FIP-0056 / CDM discussions are as follows:

Executive Summary

• On the Proposals Technicals
  • Both FIP-0056 and CDM favor Long-Duration CCs with different patterns and intensities compared to the Current Active Algorithimical Policy (Now). CDM does so heavily, while FIP-0056 does so slightly. FIP-0056 can be better for long / fully-filled FIL+ sectors and CDM for shorter / slightly-filled FIL+ sectors. FIP-0056 doesn't actively disincentivize any sector, while CDM does so: it generates a virtual penalty for having too much FIL+.
  • Both proposals can increase the amount of Locked Filecoin for 1-2 years, with CDM having a more substantial effect.
  • The Termination Fee Cap Modification in both proposals should be removed and proposed as a distinct FIP.
  • CDM can take either the proposed or an FIP-0056 like parametrization. The latter will have an more balanced QAP Share among participants and less effect on the temporarily increased Locked Filecoin .
• On the debate and general recommendations
  • The friction in the process could be improved by segmenting it according to the steps described in the Computer-Aided Governance in Action. If FIP-0056 is to be postponed/rejected, we recommend using it as the source document for the workflow around a new proposal so that there's a soft agreement between requirements, forms, and parameter values on all steps of the process.
  • The debate has exposed multiple concerns. All are legitimate in intent, but not all in factual basis. We provide a extensive review in this document.
  • FIL+ has been a focal point of disagreement among participants. We recommend that the community has space for open discussions around its role and functions and how it fits into the various Vision(s) for the Network.
  • Community Oversight of the analyses surrounding FIPs and additional Educational Efforts are to be suggested.
• On the recommendation around the Decision
  • We recommend Postponing or Rejecting FIP-0056 as per application of the Precautionary Principle on the Technological Hurdles & RBP as Key Advantage thesis as advocated by the CDM Proponents. If that thesis is taken to be Unlikely or False, then the Recommendation is Void towards FIP-0056.
  • An underlying thesis behind CDM relates to technological rather than economic hurdles being blockers to Filecoin Adoption. If Filecoin's RBP size is considered a key advantage, then there's a catastrophic risk of interfering with the CC's economic calculus. With the currently available information, we are unable to disprove that Thesis.

Diagnosis

• On the technical comparison of the FIP-0056 and CDM proposals
  • Both proposals will have long-duration CCs as their primary beneficiaries and short-duration CCs as the main negatively affected party.
  • Contrary to the discussion, FIP-0056 does not bias QAP towards FIL+. Unless CCs are disproportionally risk-averse, FIL-0056 is CC-biased compared to Now.
  • FIP-0056 will favor mostly long-duration FIL+ sectors, especially those filled with FIL+. CDM will disproportionally shift incentives from fully-filled sectors with FIL+ toward other parties. Partly filled FIL+ can benefit from CDM in several circumstances.
  • Both proposals can potentially increase Locked Filecoin Supply, with CDM being stronger than FIP-0056.
  • An incentive flaw has been encountered regarding the Termination Fee Cap in both proposals, which is even 'quirkier' in CDM.
• On the Debate
  • The debate process exposed multiple concerns. Some were legitimate in intent and factual basis, while others had honest intent but weak or even misleading factual basis.
    • Most of the latter did fall into two categories: 1) FIP-0056 skewing the Token Wealth Distribution even more towards FIL+ sectors and 2) mistaking Network-wide phenomena with Individual phenomena (e.g., lower network-wide rewards means lower rewards for the individual miner that remains / onboard all things being equal, which is not true).
    • The debate made clear that the FIL+ incentives are a central point of contention in the public debate as of now.
    • A misconception is the equivalence between FIP-0036 and FIP-0056. The latter doesn't modifies the Consensus Pledge, and it has a far gentler slope.
  • Issues around study reproducibility and perceived relevance towards stakeholders as consumers of studies were identified.
  • A concerning thesis has been propagated by the CDM proponents, which, if true, poses a catastrophic risk to the Network's Success. Given the impossibility of disproving it and the perceived non-criticality of the FIP-0056 modification, we opt to employ the Precautionary Principle for the recommendations.

Recommendations

• On the Proposal's Technical Structure
  • FIP-0056: Remove the Termination Fee Cap dependence on SDM from FIP-0056 and make it a separate FIP.
  • CDM: Consider FIP-0056 equivalent parameter choices
  • CDM: Remove the Termination Fee Cap dependence on SDM entirely.
  • CDM: Consider soft transitions rather than caps on the Quality Multiplier.
• General
  • Community Oversight on the FIP Analysis should be increased through either external stakeholder involvement during its execution or usage of independent FIP reviewers
  • Educational Content coverage on Filecoin's Economics and Protocol must increase Quantity, Reach, and Quality.
  • The community should be stimulated to discuss the Storage Market in terms of its state, challenges, benchmarks, and required actions & incentives. The inherent risks and the role of Filecoin's endogenous economy around the transition towards a Storage Marketplace / FVM-dominated world should be the topic at large.
• If FIP-0056 is not approved
  • Go back to first principles: formally describe the problem that should be solved (e.g., 'we should incentivize longer sectors,') and what mechanisms can be used to support that. Select the mechanism(s) based on diversity-of-interest criteria: the best one is the one that can be tuned to manage trade-offs across different utility functions.
  • Break down the solution proposal into two FIPs:
    • The first should decide the mechanism (e.g., 'we should have a Sigmoid SDM regardless of the numbers). This determines the Class of Incentives that are achievable.
    • The second should determine the actual numbers (parameters) that tune the mechanism. This instantiates the Actual Incentives that are to be realized.
• On Taking a Decision Immediately
  • FIP-0056 should be postponed or rejected if the "Technological Hurdles & RBP as Key Advantage thesis" validity is considered Unknown or Likely to be True.
    • FIP-0056 does not appear to be intrinsically 'bad' or 'good' for the FIL ecosystem. However, in Engineering work, the precautionary principle always applies. If the expected benefit doesn't clearly outway the potential risks, postponing is always a safer option.
    • The same precautionary principle applies to CDM as of now.
    • A possible concession is to utilize the more general proposed CDM form and set it so that it has FIP-0056 behavior: Ceiling on 20x and slope = 0.287. The definition of those two values could be the subject of a poll.
  • CDM, as being proposed now, should be modified before being considered
  • We could not find a compelling public argument making the case that FIP-0056 is an urgent intervention.

[zhiqiangxu]

Thanks for the clarification of "relative gain"!

[nai2283]

thus the relative gain is larger for DC

No the opposite, it is larger for CC. Check happens in this expression when CC goes from 1x to 2x compared to DC going from 10x to 20x?

No it's not. Doing CC is still a business of losing money even after 2x. Gas cost along is costing approximately 10 months of block reward for 32 GiB sectors, make it 5 months cost doesn't change much. And we don't even consider pledge costs like loan interests and other fiat cost here! If you are saying SDM is for CC, you will need to present your quantified ROI before and after SDM.

[tmellan]

Sure, simple example https://github.com/protocol/CryptoEconLab/blob/baseline-crossing-updated/notebooks/simple_SDM_RM_example/simple_returns_mutliple_scaling_example_python_version.ipynb

You can adapt lending and other assumptions to individual scenarios. That changes the exact numbers but not the principles, which are:

• Compared to today: SDM is expected to make committed capacity (CC) more viable.
• Comparing CC to Verified Deals: the possible relative gain in the Returns Multiple (rather than just bare multiplier), is larger for CC.
• Verified Deals are expected to remain the best option.

[zhiqiangxu]

@tmellan is there a spec on how Fil+ actually do the verification? Open and public is very important for Fil+ to be accepted by the network.

[tmellan]

Hey @zhiqiangxu,

Notaries act as fiduciaries, the process and guidelines are described here and here. In terms of more operational nuts and bolts and how Fil+ is evolving, @dkkapur is the probably the best person to advise.

[nj-steve]

@dkkapur
Hello, everybody! I‘m steve from Antalpha.We recently did a survey of SPs and Clients about FIP-56.

• Clients:
  As enterprise clients, they have the requirement that the maximum sector duration will increase from 1.5 years to 5 years.
  Because some of them want to keep their data in filecoin for more years, they don't need to store it again on their own servers.

• SPs:

1. Some SPs said they supported the proposal, and expected the FIP in next fork.But they don't want to store too long,Because it will takes 20X as much to pledge when they increase the duration to 5 years.If the pledge will take 10X and SPs get 20X power, so even if the duration is 5 years,they will be very pleasure to increase the duration to 5 years.
2. Some of them were lukewarm about the proposal.They need to release the pledge to pay to someone else.
3. A part of them think the FIP is acceptable.They can save time when the sectors come to expiration but they want to continue sealing.
4. Some people don't like the FIP,they think it's just an economic behavior.

[dkkapur]

thanks @nj-steve! Can you share some more details about who the SPs and the Clients were, even if it is in terms of how and where you ran the survey? How many participants?

Thanks for sharing the outcomes!

[nj-steve]

Hello @dkkapur We ran the survey inside our lab wechat group about 299 users. We also called to a few SPs and investors and did the research. About 21 people participated in the discussion and gave feedback.They are in North America and Asia.

[dcasem]

I would like to point out that we are starting to see a fairly significant decline in block rewards, even with FIL+ today. As it was explained to me by @flyworker, this is because the baseline is predominantly based on CC; if raw power lower is lower than the baseline, block rewards decrease.

I'm a bit outside my depth here, so would love to hear opinions on this from @SBudo, @momack2, @tmellan and others on how FIL-56 or the CDM alternative proposed may have an impact to block rewards going forward.

[zixuanzh]

Thanks for the comments @dcasem @SBudo , we have noted the trend of declining network raw byte power and the impact of baseline crossing many months ago and been monitoring the situation with open sourced model and reports. You can find them on CEL's website and almanac. This is part of why we started the discussion and proposal of duration incentives since June 2022 to encourage longer commitments. Many of the macro trends that we worried about have since materialized which is not great for the network.

FIP36 was proposed as a force to counter that downward pressure (expiring sectors, declining network raw byte power, etc) on the network a year ago. Community then was vocal against it because it was perceived to be "supporting CC". Circumstances have changed, FIP56 is also different from FIP36 and FIP56 is more focused on encouraging longer duration (which contributes to the slowing down of raw byte power) than some other improvements in FIP36. FIP56, as @danlessa pointed out during the AMA, is pretty mild but it's still helpful.

CEL has looked into and analyzed CDM. As currently specified, CDM will likely not cause more CC to join the network and may accelerate the decline of raw byte power. There is also more detailed analysis above. However, we recognize that the network wants more hardware commitment, not only for block reward's sake but also for the strength of our ecosystem. Teams are also actively working on other solutions that entice more hardware commitment while also bringing positive changes to the network. You can find the discussion here #666. Nonetheless, regardless of the solution we arrive at the entice more hardware, that solution works better with a duration incentive for the network.

[herrehesse]

@zixuanzh You have stated:

“CEL has looked into and analyzed CDM. As currently specified, CDM will likely not cause more CC to join the network and may accelerate the decline of raw byte power.”

Not only does it contradict the views of the entire active SP community, but it also lacks clear evidence or logical reasoning. Hence, I kindly request that you provide additional details and include the justification for why SDM performs better.

In my perspective, this assertion is unfounded.

[tmellan]

Hey @dcasem , thanks for asking the question.

@zixuanzh has summed it up, but let me add a few more details on your point:

how FIL-56 or the CDM alternative proposed may have an impact to block rewards going forward.

General

• @flyworker is correct, CC makes up more of the raw byte power at the moment.
• If onboarding and renewal rates increase leading raw byte to increase, then the baseline component of the minting will increase. And vice versa.

FIP-0056

• If we maintain the same onboarding and renewal rate, and if the policy is effective in increasing average duration, then minting can go up compared to the current trajectory (with no FIP implemented). The reason is if expirations are lagged this can boost raw byte.
• If we maintain the same onboarding and renewal rate, and if the policy is _in_effective (average duration stays the same as now), then the minting will be no different to the current trajectory.

Risks to consider

• If we were in a world where the pie is fixed in size, then changing incentives to encourage duration means something has give somewhere else. That something could be the level of CC onboarding or it could be the level of Fil+ power being added.
• This is what we want to avoid and is the key argument behind applying the Precautionary Principle. Broadly speaking, there are two approaches to apply this: (i) postpone/reject the FIP, (ii) mitigate the risk, by finding a parameterisation of the policy with minimal scope to cause harm. The second is the approach we’ve taken by changing to the gentle slope of 0.285 (max multiplier of 2x at 5y).
• Important note: this is all considering an if pessimistic scenario framing with fixed pie size. This isn’t what will happen but the point is to consider all possibilities especially the negative ones. And by all accounts, the potential to grow the onboarding is very substantial in the coming year, with the many innovations already here/incoming that were reported on Thursday: FVM enabled lending, improvements to Fil+, sealing cost improvements, possible pledge shortfall FIP.

Comment on CDM and SDM differences

• The issue with CDM is that in the pessimistic scenario mentioned above, it has the potential to make it a whole lot worse because the maximum multiplier from duration is ~8.5x. This means CC could get the same power with x8.5 less hardware. And therefore lead to lower raw byte level and so minting.
• Would this actually happen with CDM? Probably not to the x8.5 extent for a number of reasons: (i) CC opting for the max duration to get the max multiplier is unlikely given the long duration commitment involved, (ii) for current CC, it depends on hardware resale costs and inertia to change business model, (iii) through cannabilizing verified deals, CC could have sufficient growth to outweigh the consolidation effect, albeit at the cost of verified deals, (iv) with all the innovations coming along there’s substantial scope for growth generally.
• Overall though, the risk is sufficient that it can’t be dismissed. Which is why based on the same logic, FIP-0056 opts for a much weaker set of changes.
• Can FIP-0056 increase raw byte compared to the current trajectory? Potentially, through (i) giving CC the option to be more competitive and grow, slightly closing the gap between Fil+ and CC. (ii) Creating delayed expirations, which have the potential to support network raw byte level. But this also depends on the factors set out above.

Appreciate this is not the simplest answer, but wanted to set out some of the nuances. Happy to give more clarifications.

[momack2]

The above explanation from @tmellan is more thorough - but I wanted to give a really short answer to @herrehesse's question:

“CEL has looked into and analyzed CDM. As currently specified, CDM will likely not cause more CC to join the network and may accelerate the decline of raw byte power.”
Not only does it contradict the views of the entire active SP community, but it also lacks clear evidence or logical reasoning. Hence, I kindly request that you provide additional details and include the justification for why SDM performs better.

• CDM creates a strong incentive for capital-constrained SPs to aggregate their current sectors into less hardware - accelerating raw byte power (RBP) decline.
  • AFAIK all/most SPs express access to collateral as their main challenge to grow right now
  • Holding collateral static, an SP is incentivized to aggregate their current sectors into less hardware (aka less RBP), maintaining their same network power ratio with less hardware expenses and putting even more downward pressure on network growth.
  • That is because the up-to~10x multiplier for CC sectors, which allows up to ~10x consolidation while maintaining the same collateral and power.
• SDM, which only offers a 2x multiplier for both CC and FIL+, effectively guards against this potential 10x RBP drop while (hopefully) increasing CC renewals and longer commitments to mitigate RBP decline.
  • The CEL modeling (ex here and extreme cases here) highlights that SDM with this shallower slope creates an incentive for participating CC SPs to renew their existing sectors and commit them to the network for longer, without creating significant incentives to terminate / consolidate sectors. This is good for everyone.
  • SDM is also more accessible to CC SPs (who aren't limited on sealing new FIL+ deals), and therefore creates an upside for renewal and longer commitments of existing RBP within the network (mitigating the current decline and creating a positive impact on RBP and CC sector profitability).

Honestly, the fact that some SPs claim to be unsupportive of landing SDM ASAP is nuts to me - because it mitigates the exact same concerns that folks claim to desire out of CDM (additional upside for CC SPs, incentives to decrease RBP decline, positive incentives for longer sector and deal commitment, etc). We can and absolutely should do additional improvements to cryptoecon in the future, but IMHO it's a tragedy to throw out a valuable mitigation and network improvement that's ready NOW, due to misinformation and misunderstanding. 😓

I found the deeper analyses and AMAs between CEL, Blockscience, and others very clarifying to help highlight and correct some of these misconceptions. Did you catch the AMA last week (https://www.youtube.com/watch?v=F46MeZfy7cw)? Did that help clarify your understanding?

[herrehesse]

In my previous message, I discussed the issues I have with the proposed SDM protocol update for the Filecoin network and why we would vote for CDM. However, I feel it is also important to highlight an additional challenge that every storage provider faces, namely the challenge of obtaining Filecoin+ datacap. It is worth reminding everyone that obtaining Filecoin+ datacap has been the sole means of survival for storage providers on the Filecoin network for over 11 months now.

As you are no doubt aware, the Filecoin+ program has been plagued by fraudulent & abusive activity, with many entities attempting to exploit the program to generate revenue without actually providing valuable storage services. This has resulted in a collapse in revenue for legitimate storage providers, as well as a rise in the number of abusive entities seeking to profit from the program.

Therefore, I must strongly object to your proposal to increase the possible multiplier for FIL+ sectors to x20. Given the current state of the Filecoin+ program, increasing the multiplier would only exacerbate the current problems the program faces. Rather than adding more multipliers to a challenged ecosystem, we should focus on resolving the underlying issues and improving the overall health of the ecosystem.

Although CDM recognises numerous reasons for the creation of SDM, it advocates for maintaining the static 10x multiplier on Filecoin+ as it appears to be the most reasonable approach considering the current state of the program.

In conclusion, I urge you to reconsider your undoubting support for the proposed SDM upgrade and to take into account the concerns of the wider community, including many well-established and active storage providers who share my perspective. By adding a x2 to CC none of the issues are addressed.

Adding: People's opinion about something that is not in line with your believes should never be considered as "nuts".

[Stebalien]

The current termination fee cap modification can be easily circumvented by either:

1. Extending the sector's lifetime by six months, if the remaining sector lifetime is less than 1 year and the total lifetime is less than 4.5 years.
2. Otherwise, by snapping a small deal to reset the power base epoch.

[vkalghatgi]

Got it: I think I remember discussing something along these lines with @ZenGround0 when initially prototyping the FIP. Instead of keeping a running total of termination penalties, can this be mitigated by using the maximum sector duration when calculating fees? So, when calling pledge_penalty_for_termination which should now have a duration field to calculate termination fee, use duration := sector.expiration_epoch - sector.activation_epoch

[Stebalien]

That works, but that would apply the extra penalty to any long-lived sector even if it has never benefited from SDM.

[vkalghatgi]

That's true. Alternatively, I think you could let duration := max(sector.expiration_epoch - sector.power_base_epoch, sector.base_power_epoch - sector.activation

[Stebalien]

Slightly better, but it still has the same issue for a long-lived sector.

[Stebalien]

Really, expiration - activation sounds like a reasonable solution: it makes the penalty grow (somewhat) proportionally to the lifetime of the sector. But we need to recognize that as a deliberate choice, write it into the FIP, and argue why it makes sense.

My primary concern is that we've tossed in too many "hacks" to make this FIP work:

• Resetting the power base epoch (and SDM) on snap deals (potentially reducing power, I'm still worried about this one but couldn't get any response).
• Keeping the maximum replaced day reward and pledge penalty.
• And this.

These kinds of things:

1. Make it painful for users. E.g., why am I being penalized X for Y when it should be Z.
2. Make it difficult to maintain the protocol.
3. Need to be discussed as a part of the FIP, not treated as "implementation details".