Skip to content
New issue

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

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

Already on GitHub? Sign in to your account

Draft of Removing Restriction to Access to Multiplier and Allowing Unrestricted Minting of Datacap #803

Merged
merged 38 commits into from
Oct 11, 2023
Merged
Changes from 8 commits
Commits
Show all changes
38 commits
Select commit Hold shift + click to select a range
45ea13b
Draft of Phasing Out Fil+ and Restoring Deal Quality Multiplier to 1x
dcasem Aug 16, 2023
5ecbbde
Update fip-0071.md
dcasem Aug 16, 2023
7a96cbb
Update and rename fip-0071.md to FIL+ FIP.md
kaitlin-beegle Aug 16, 2023
1553ffa
Create draft-fcm-fip.md
dcasem Aug 23, 2023
aafbb9c
Delete fip-0072.md
dcasem Aug 23, 2023
272e459
Draft of Removing Restriction to Access to Multiplier and Allowing Un…
dcasem Aug 23, 2023
3f57e08
Delete FIL+ FIP.md
dcasem Aug 23, 2023
a0a2046
Update fip-FCM.md
dcasem Aug 23, 2023
9755f3c
Update fip-FCM.md
The-Wayvy Sep 4, 2023
3293dd5
Update fip-FCM.md
The-Wayvy Sep 4, 2023
05ba49f
Update FIPS/fip-FCM.md
The-Wayvy Sep 5, 2023
aaa8481
Update fip-FCM.md
The-Wayvy Sep 5, 2023
4be70eb
Update fip-FCM.md
The-Wayvy Sep 6, 2023
5ca56a9
Update fip-FCM.md
The-Wayvy Sep 6, 2023
17086bc
Update fip-FCM.md
The-Wayvy Sep 8, 2023
1c04026
Update fip-FCM.md
The-Wayvy Sep 13, 2023
ceef055
Update fip-FCM.md
The-Wayvy Sep 14, 2023
73724fb
Update fip-FCM.md
The-Wayvy Sep 14, 2023
8d378c6
Update fip-FCM.md
The-Wayvy Sep 20, 2023
9bc27ed
Update fip-FCM.md
The-Wayvy Sep 20, 2023
729a99b
Update fip-FCM.md
The-Wayvy Sep 20, 2023
66d4820
Update fip-FCM.md
The-Wayvy Sep 20, 2023
f5bf466
Update fip-FCM.md
The-Wayvy Sep 21, 2023
6af1eeb
Update fip-FCM.md
The-Wayvy Sep 21, 2023
053a4b8
New discussion forum link
anorth Oct 3, 2023
c1694b7
Update and rename fip-FCM.md to fip-0078.md
The-Wayvy Oct 4, 2023
769c4ed
Update README.md
The-Wayvy Oct 4, 2023
bb68568
Update README.md
The-Wayvy Oct 4, 2023
3659ada
Update fip-0078.md
The-Wayvy Oct 4, 2023
f8e2597
Fix header delimiter
anorth Oct 4, 2023
953aca3
Merge branch 'master' into dcasem-patch-1
anorth Oct 4, 2023
2411bee
Update FIPS/fip-0078.md
The-Wayvy Oct 7, 2023
57d94f9
Update FIPS/fip-0078.md
The-Wayvy Oct 7, 2023
1d5af53
Update FIPS/fip-0078.md
The-Wayvy Oct 7, 2023
fe97e1a
Update FIPS/fip-0078.md
The-Wayvy Oct 7, 2023
7ca601d
Update FIPS/fip-0078.md
The-Wayvy Oct 7, 2023
accb618
Update FIPS/fip-0078.md
The-Wayvy Oct 7, 2023
f4f8dd7
Update FIPS/fip-0078.md
The-Wayvy Oct 7, 2023
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
72 changes: 72 additions & 0 deletions FIPS/fip-FCM.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,72 @@
```---
fip: "FIP-XXXX"
title: "Removing Restriction to Access to Multiplier and Allowing Unrestricted Minting of Datacap"
anorth marked this conversation as resolved.
Show resolved Hide resolved
author: Fatman13 (@Fatman13), flyworker (@flyworker), stuberman (@stuberman), Eliovp (@Eliovp), dcasem (@dcasem), and The-Wayvy (@The-Wayvy)
Discussions-to: https://github.com/filecoin-project/FIPs/discussions/774
Copy link
Member

@jennijuju jennijuju Aug 30, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I would suggest we create a new FIP discussion just for this proposed solution so that we can have a follow-up discussion on this specific proposal in a less clustered place.

Copy link

@The-Wayvy The-Wayvy Aug 30, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Create a new discussion, miss the deadline, and then have to go through another 6 months of permissioned nonsense?

No.

CEL has already done the analysis and rubber-stamped this solution.

This FIP also does not close the door on Filecoin Plus 2.0

We need fair, transparent and objective block reward distribution.

The notary system has proved itself to be a failure in practice and has always been a failure theoretically because it nerfs the exact properties that PoSt/PoS/PoW enabled.

We've already spent months talking about this.

The notary system has probably onboarded less than 10 PB of paid data storage, and in the process, we have disenfranchised the majority of the network's miners and gotten everyone at each other's throats.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Jennifer is requesting a new discussion so that folks who read this proposal have a place to go comment and discuss it going forward that isn't already crowded with many competing/alternate proposals that are different to the one described here. This enables clearer discussion and debate with a new draft proposal.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I agree with @jennijuju and @momack2 here and have created #844. This should not be taken as any rewinding of process, but a venue for discussion of this concrete proposal to aid a quicker governance process.

status: Draft
type: Technical
category: Core
created: 2023-08-22
requires:
replaces:
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please specify that this would need to replace FIP0003.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sure

```

## Simple Summary
This proposal aims to remove restrictions associated with the 10X multiplier in Fil+ deals and allow any actor to mint Datacap, promoting decentralization and simplifying the Filecoin ecosystem.
Copy link
Member

@jennijuju jennijuju Aug 30, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The title suggests "Removing Restriction to Access to Multiplier" & "Remove restrictions associated with the 10X multiplier in Fil+ deals" - what does this mean?

are you suggesting who may call AddVerifier and AddVerifiedClient, or you simply meant who can mint Datacap -> which is the only change i saw proposed in the spec

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't think any change to AddVerifier/Client is needed to accomplish the authors goal here. Those only serve to funnel the datacap minting process through the verifreg. Am I missing something?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I also dont think they are needed - thus im confused with the phrasing of "Removing restriction to access to multiplier" in the title - to me it implies a change anyone can deploy their own multiplier and gets additional QAP thats other than the deal quality multiplier that comes with datacap.

That being said, this proposal should also call out the degraded functionality and/or the unnecessity of AddVerifier and AddVerifiedClient methods, with the introduction of "all you can mint" datacap rule.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@jennijuju

You are the expert on implementation.

We appreciate your help.

anorth marked this conversation as resolved.
Show resolved Hide resolved
anorth marked this conversation as resolved.
Show resolved Hide resolved
anorth marked this conversation as resolved.
Show resolved Hide resolved
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
This proposal aims to remove restrictions associated with the 10X multiplier in Fil+ deals and allow any actor to mint Datacap, promoting decentralization and simplifying the Filecoin ecosystem.
This proposal aims to allow any actor to mint Datacap and bestow a 10X multiplier to any sector, regardless of whether it stores client data / serves FIL+ deals.

Seems like this is a more accurate and descriptive summary for this proposal

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Agree with this edit, it is more accurate than unsubstantiated commentary around decentralization, simplification, or permissionlessness.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@galen-mcandrew

Our commentary on permissionlessness is pure fact.


## Abstract
The Fil+ program was designed to encourage high-quality storage on the Filecoin network. The current proposal intends to remove complexities that have arisen from its initial conception, thereby ensuring the Filecoin mission is better represented and uncomplicated.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

High-quality storage sounds ambiguous to me, as one can also provide "high quality" committed capacity storage on Filecoin.

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

^^ I think a lot of this language is ambiguous.

If the simple summary gives a 1-2 sentence overview of the changes you want, the abstract should add a bit more detail about why. Saying that you want to "remove complexities" and "ensure the Filecoin mission is better represented" is a unclear. Just say what the complexities are, and how your proposed changes service the Filecoin mission.

Re: @jennijuju's points above, one definition that you can use for Fil+ is the one in FIP0003: "... maximize the amount of useful storage that Filecoin can and will support."

I want to help you avoid vagueness as much as possible. If you have vague language and terms in your FIP draft, it makes it harder to navigate the soft consensus/last call process.

Copy link

@The-Wayvy The-Wayvy Aug 30, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I would replace

The current proposal intends to remove complexities that have arisen from its initial conception, thereby ensuring the Filecoin mission is better represented and uncomplicated.

with

The current proposal intends to remove the permissioned notary system, ensuring that Filecoin remains decentralized and fair.the current proposal intends to remove the permissioned notary system, ensuring that Filecoin remains decentralized and fair.

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No, these are the same things. Words like "permissioned", "decentralized", and "fair" don't clearly frame your proposal. You need to say 1) what concrete change you're making, and 2) for what specific, concrete reason. Adding vague words- especially ones like "decentralized", which can be interpreted and defined in many different ways- makes both governance and implementation harder to navigate.

So, for example:

"The Fil+ program was designed to encourage high-quality storage on the Filecoin network. However, [insert specific problem you're trying to address]. To address this, we propose a change to the maintenance of datacap, such that datacap can be accessed by any network participant. Doing so will enable us to solve [insert aforementioned problem] because [insert reason]."

In other words, you're not making an argument; you're specifying a proposal. Does that make sense?


## Change Motivation≠
The current Fil+ implementation, though well-intentioned, has added complexities that deviate from Filecoin's core mission. Simplifying this system will encourage the onboarding of genuine data and uphold the integrity of the Filecoin blockchain.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you elaborate on the complexities and the issue/challenges associated with them? (They were discussed in the discussed, lets summarize the points that your proposal can resolve. We should ensure FIP readers may get enough context without going through the whole fip discussion)

Copy link

@The-Wayvy The-Wayvy Aug 30, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The motivation for this FIP is less the complexities of FIL+ and more that it’s permissioned.

We talked about this at length in 774.

For new money to come into the ecosystem, be it VCs or Miners, they must believe that this is a fair, transparent chain that uses a consensus algorithm grounded in objective, verifiable computation.

There is no demand for permissioned blockchains. The market doesn't want that and never has. It isn't a good design and it does not incentivize useful storage.
At best it misallocates subsidies to users who are unwilling to pay for services and at worst it's a complete charade.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I would replace

The current Fil+ implementation, though well-intentioned, has added complexities that deviate from Filecoin's core mission. Simplifying this system will encourage the onboarding of genuine data and uphold the integrity of the Filecoin blockchain.

with

The current Fil+ implementation, though well-intentioned, has created an uneven playing field for SPs, permissioned the chain, and encouraged undesirable behaviors like bribery and false accusations. Our proposal makes Filecoin fully permissionless. It will remove these problems and restore confidence in the integrity and neutrality of our consensus algorithm

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you please explain how this proposal will "encourage the onboarding of genuine data"? This seems core to your argument, but is unsubstantiated in this proposal.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I would like to share that there is no concrete evidence of bribery, even though multiple claims have been made about secondary markets for DC. Additionally, all accusations, as they concern abuse of DC, are well founded in evidence here - https://www.notion.so/filecoin/T-T-Dispute-Tracker-d28b93677cb544b48e77b585856601cf

Copy link

@The-Wayvy The-Wayvy Sep 4, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@cryptowhizzard

You are a notary and do not want to lose your privilege.
That is why you defend the current system.

@hyunmoon

I am with you that we need paying users for the network to succeed.

But there's a right way and a wrong way to go about promoting paid deals.

The notary system is clearly not the way.

Copy link

@The-Wayvy The-Wayvy Sep 4, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@cryptowhizzard

This is really strange. Fil+ is about onboarding valuable data and onboarding this takes effort, resources and time.

The question is: who should decide what is valuable?

I believe users should decide what is valuable and what is a good use-case for our technology.

The governance team and notaries are not the ones paying for services nor are they using their own money to subsidize deals. Therefore they should not be making decisions about what is quality and useful.

Having non-users dictate 'quality' and 'utililty' is speculative at best and corrupt at worst.

If you want to onboard crap and trash I think people are welcome to do that, just not with 10x.

Again, data you don't like is not 'crap'

The usefulness of data = how much users will pay for it

You don't know what Filecoin's PMF will be.

This is like saying Uniswap is crap because you can't trade 'real assets' like US Treasury Bonds and NASDAQ stocks.

And why is a permissioned system bad? If the goal is onboarding true and real data there is nothing wrong with it. The argument “no market” makes no sense at all, there is market and there are customers atm. If those customers don’t want have anything to do with company’s who are not suiting them well for their needs it is up to them to change and make themselves competitive.

The whole point of having a sybil resistance mechanism on a network is to enable permissionless participation.
PoSt and PoS are sybil resistance mechanisms.
If you want a permissioned network then there is no need to waste so much capital in PoS and so much computation in PoSt.
We can KYC every storage provider and have them sign contracts with PL.

Literally the entire point of this technology is to enable permissionlessness.

2

'True' and 'Real' data does nothing for the price of Filecoin is users aren't paying for it.

Marketing slogans are no substitute for Product Market Fit.

3
If you want a fair protocol then at the very least you should demand that all Filecoin Plus deals go up for public auction.

Copy link

@The-Wayvy The-Wayvy Sep 4, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@hyunmoon

If the governance team / notaries want to play VC they can use their own money to subsidize SPs serving 'real' users.

Their speculation/preferences are not a legitimate input to a consensus algorithm.

Copy link

@hyunmoon hyunmoon Sep 4, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I concur. Ideally, the project should have been backed by a single, upfront allocation of funds. However, I find it challenging to accept that the Filecoin Plus system can operate effectively as a fully permissionless entity. If there were a practical way to achieve this, I'm confident that the adept team at Protocol Labs would have already executed it.

I contend that the issues might not have escalated to this extent if the project had remained under the management of Protocol Labs employees instead of transitioning to elected notaries. The project ought to have been overseen by the very entity footing the bill. Although the initial vision might have centered on community-driven funding and governance, it appears the outcomes have fallen short of expectations.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I addressed how 803 will disincentivize the onboarding of data with no paying users and fabricated datasets in "Product Considerations"


## Specification
**Change in the Datacap Actor's Mint Function**: Alter the DataCap actor's mint() function to remove the restriction that the caller must be the "governor" (the verified registry actor). Allow any actor to mint datacap.

## Design Rationale
While more comprehensive solutions might be considered in the future, this FIP offers an immediate, simpler approach.

Allowing any actor to mint datacap is the simplest change to the current protocol that removes the Fil+ process restrictions on datacap issuance. This datacap can then be transferred to the verified registry to create a verified allocation and claim the associated multiplier when onboarding data.
The-Wayvy marked this conversation as resolved.
Show resolved Hide resolved
After this change, the verified registry and datacap actors will be mostly obsolete. Allowing unlimited datacap does not create any complexity, but does make it possible to remove some. Future protocol improvements could remove these actors and support direct specification of a multiplier when onboarding to further simplify the protocol and reduce cost.
anorth marked this conversation as resolved.
Show resolved Hide resolved

## Backwards Compatibility
This proposal does not change any chain state or method interfaces so is backwards compatible to all clients. A network upgrade is required to deploy the new actor code.

After this upgrade, Storage Providers (SPs) can engage with any deal, whether Fil+ or not by minting Datacap. Deals will still need the `verified` flag set to true, but clients will be able to freely do this because they can mint the required datacap.
anorth marked this conversation as resolved.
Show resolved Hide resolved


## Test Cases
Test that an account actor that is not a notary can mint datacap
Test that a non-account actor can mint datacap

## Security Considerations
This FIP, though providing simplified access, doesn't introduce new security threats.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we will need analysis on this -what is the cost for any/group of malicious actors to mint large amount of the data cap and onboarding CC sectors with the 10x multiplier and get>30% of the network QAP.

Copy link

@ArthurWang1255 ArthurWang1255 Aug 30, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The security of the network will be increased by the fip. In the past, only 1/10 of the hardware + enough DC was needed to do evil. Now it needs 1:1 hardware to be evil.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I agree with Jennifer that this section seems incomplete. Security consideration should be given to:

  • could a large CC-sector miner use this change to quickly upgrade all their CC sectors into fake "10x deal" sectors (without actually onboarding any real/useful client data) and use that to gain control of the network?
  • how does this change harm the security of data on the network? (ex it significantly decreases the confidence of existing Filecoin clients that SPs will continue to store and safeguard their data - since they could instead dump their sectors full of real client data and get the same protocol incentive for storing strings of 0s - which are easier to repair and come with no client expectations)

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Currently, DataCap has no 'expiration point' and can only be destroyed by notary and RKH message (per FIP0028) or burned in deals.
Are there additional concerns of actors minting massive amounts of DataCap? What would happen if a bot opened new client addresses and spammed that mint DataCap function, creating EiB of DataCap immediately?
Should there be a threshold or gate on minting?

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What's the difference between fil+casting datacap and free casting datacap, especially given the amount of abuse?
What I want to know is why you don’t need to consider security issues when you turn on fil+ casting datacap, but you need to turn on free casting?

Is it more straightforward to switch to staking mining (POS) in one step?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Regarding the risk of rapid gains in power from upgrading CC, thanks for the napkin analysis but the point of the question is the FIP document must discuss this the presence and likely size of this risk, and any mitigating factors. You might also note that SPs would have to snap-deal to do that upgrade, which introduces a hardware constraint. But again, the point isn't to convince anyone here, the point is that the document needs to contain a reasonable discussion of the risk.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Re existing client data, I don't really think it's a security consideration, but it's definitely an incentive consideration. The change in incentives to SPs currently storing verified data is real. I tend to think the practical impact is low since "dumping" the data requires paying termination fees and re-sealing. But again, the FIP document must lay this out.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If this FIP doesn't discuss data security of client data here, then it needs to be explicitly discussed in the product section - both for new data onboarded (how secure is that data vs status quo) and data that's already been onboarded into the network. IMHO is appropriate to include in the security section because ensuring storage is not suddenly dropped or storage quality reduced is a security concern for achieving the core purpose of the Filecoin network - but as long as it's included clearly it could be in either section

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I briefly discussed this in a PR just now.

Copy link

@The-Wayvy The-Wayvy Sep 21, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I addressed the risk of a substitution attack in "Security Considerations"
and dumping risk in "Incentive Considerations"


## Incentive Considerations
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I still have some hesitation about the doneness of this proposal, and I think it has to do with this section.

@jennijuju and I each flagged above that we'd like more specific language, specifically where "complexities" and "simplification" are referenced.

I'd also like to see a bit more documentation about the assumptions/observations driving this desired change, and what it actually means for anyone to be able to mint datacap. I know what they are, and they've been well discussed by folks engaging on this discussion topic. Those details should at least be summarized here as well (or below in Product Considerations; doesn't matter too much).

This FIP should not just direct implementers, but should also justify and explain the policy. It would be great for you to add a few sentences about how those minting datacap will (presumably) be opting for higher collateral rates, and why they may choose to do so. Why would people choose to mint datacap if its otherwise available to everyone? Or rather, why would someone NOT mint datacap if it's giving them a 10x multiplier? Etc.

I know you know these things; they should be here too.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@kaitlin-beegle

if you add me as a collaborator I will speed up the editing process.

This was not David's first choice and he just wanted to get the PR out the door, not write a masterpiece.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@The-Wayvy we have no ability to give you edit access here since the source of the PR is a repository under @dcasem's GitHub account. He should be able to do so, though.

The proposed change effectively removes a significant incentive for the onboarding of data that was previously determined as valuable by the off-chain notaries under the Fil+ mechanism. As a result, it's anticipated that the overall rate of data onboarding may decrease. The extent to which the currently incentivized data under this mechanism is fraudulent remains uncertain.
anorth marked this conversation as resolved.
Show resolved Hide resolved

With the removal of the restriction, effectively-empty sectors can now benefit from a multiplier. This adjustment allows Storage Providers (SPs) to onboard more Raw Byte Power (RPB) at a potential profit, potentially leading to an increase in such sectors being onboarded.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
With the removal of the restriction, effectively-empty sectors can now benefit from a multiplier. This adjustment allows Storage Providers (SPs) to onboard more Raw Byte Power (RPB) at a potential profit, potentially leading to an increase in such sectors being onboarded.
With the removal of the restriction, effectively-empty (a.k.a committed capacity) sectors can now benefit from a multiplier. This adjustment allows Storage Providers (SPs) to onboard more Raw Byte Power (RPB) at a potential profit, potentially leading to an increase in such sectors being onboarded.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

With the removal of the restriction, effectively-empty sectors can now benefit from a multiplier. This adjustment allows Storage Providers (SPs) to onboard more Raw Byte Power (RPB) at a potential profit, potentially leading to an increase in such sectors being onboarded.

please elaborate what are the incentives that will bring to the network/SP by enabling onboard more CC/RBP and drive towards potential profit.

Copy link
Member

@anorth anorth Aug 30, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@jennijuju they won't be CC sectors. They will have deals in them because that's how to onboard the datacap.

Copy link
Member

@anorth anorth Aug 30, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This would actually be a strong motivation to follow on with the clean-up work of removing datacap and verifreg if this FIP is accepted. And implementing re-snap to support re-use of the capacity.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The incentive to onboard is just block rewards, the same incentive currently for onboarding 10x sectors today.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@momack2
@galen-mcandrew

Crypto-Economic Analysis is a job for... Crypto-Econ Lab

They've already done the work.

https://hackmd.io/zp-40inqSny5HMqmTSOKxw?view

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This substitution risk similar to FIP56 is important to explain in the FIP document. You can note that it's similar to risk of such behaviour if FIL+ was significantly more efficient and so datacap easier to get (even if all "good" data). It won't change the behaviour of SPs currently getting FIL+ much, which account for most onboarding today. It could change behaviour of others, but it's rather hard to say in which direction: balancing the substitution idea is the idea that for SPs who can't/won't get FIL+, onboarding anything at all becomes profitable, an improvement to today.

This risk comes down to the ability of SPs to chose their hardware/data capital balance. I tend to this this could be good for SP business models (heterogeneous environments and setups can optimise for their comparative advantage) but it does introduce the risk of this substitution having an impact on RPB. The FIP must discuss this.

It is not sufficient to say anything is someone else's job. This FIP document must describe all the relevant considerations. You can take some things from CEL's existing analysis. If more is needed you can ask them to help you, and they probably will. But no-one is obligated to do the work for you.

Copy link
Contributor

@Fatman13 Fatman13 Sep 8, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

How about the following edits regarding substitution risk? @anorth

There may be different potential outcomes with regard to substitution risk, which refers to a CC SP who may expire its RBP and use the released collateral to achieve same amount of QAP power as before but with less hardware. In the scenario where we don't have direct data on-boarding, the network growth would still be limited by the chain throughput of PublishStorageDeal messages and gas cost associated with the messages which in turn minimizing substitution risk and effectively maintaining the current network status quo of RBP falling. In the scenario where we do have direct data on-boarding, substitution risk would depend on how fast different SPs could react to the new FIP, recalibrate their hardware deployment to maximize their data on-boarding efficiency (hardware/data capital balance) and new chain throughput of direct data on-boarding message. Substitution risk in second scenario is also mitigated by the fact that CC SPs are expiring sectors anyway given the history CC expiring trend. Substitution in such case may even help to retain the SPs to stay in the network longer until the chain reaches equilibrium again.

Copy link

@The-Wayvy The-Wayvy Sep 13, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@momack2

We would like to cut the multiplier to keep RBP from further contracting.

One variation on 803 would be to give everyone access to the multiplier, but also gradually cut it down to one.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Resolved


Following this amendment, primary motivations for SPs to onboard data will still need to originate from outside the core protocol. This could encompass direct client payments or potential incentive models established within upcoming user-defined smart contracts.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you give a rough estimate on the drop in current inbound client demand for Filecoin that would drop off given the scale required for these "direct client payments", either here or in the product consideration section?

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I edited the FIP to mention that drops in demand are hard to quantify because
1.) We don't know what % of existing data on Filecoin has a real user behind it.
2.) We don't know what users are actually willing to pay for Filecoin services since they've been getting storage for free on Filecoin Plus.



## Product Considerations
Clients and Storage Providers will experience a streamlined interaction with the Filecoin network, which could lead to more user-friendly products and services built on top of Filecoin.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you please clarify why you expect more user-friendly product and services to exist without client data storage incentives?

Copy link

@The-Wayvy The-Wayvy Sep 4, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Without the notary system there would be fewer steps to getting their data onboarded.

Storage incentives are fine, but they can't have a human in the loop, and all computation involved needs to be verifiable by miners.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

there already exists pathways to onboard data without FIL+ incentives or using the notary program - however we have not seen any user-friendly products or services building around that data onboarding flow (that I know of - maybe Telnyx offers one that is publicly accessible?). I don't think you have evidence to support your claim that better UX will just happen without economic incentives (ex, multiple data onboarding solutions - offering much better UX) have been created around the FIL+ data onboarding flow including Spade, Singularity, web3.storage, BDE, Motion, and many tools created directly by storage providers.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I changed it to:

Removing the notary system will streamline data onboarding for users as well as lower costs and uncertainty for miners.

Miners are still free to use all those data onboarding tools that you mentioned.

Under this proposal, the only economic incentive for miners to improve data-onboarding UX would be to get more paying users.

We think that's a good thing.

I added it to incentive considerations.


Fil+ team and many more customized deal markets can be developed on L2 of Filecoin in either permissioned or permissionless manners in an open and fair market competition.

The proposed changes present SPs with a more flexible environment, allowing them to select a pledge-to-hardware ratio best aligned with the specific costs and benefits they encounter. This enhanced flexibility means SPs can make more rational decisions based on their individual circumstances.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Could you please add how would the proposed change impact the data client?

Copy link

@The-Wayvy The-Wayvy Aug 30, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The SPs are the real users of the network for now because they pay for the vast, vast majority of Filecoin blockspace / transaction fees.

Data clients are a negligible constituency at the moment.

Copy link

@ArthurWang1255 ArthurWang1255 Aug 30, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Agree with @The-Wayvy . Filecoin is a blockchain, and the users of a blockchain are those who hold and consume Filecoin through gas. Unless data clients directly pay fil to store on Fileocin, the impact to data clients are not directly related to the filecoin core protocol. IMHO, the fertility of RBP and SPs is what needs to be considered in a fip.

But from a platform perspective, if the fip can bring more RBP and SPs, it will generally bring more data clients and let them receive better services.

Copy link

@The-Wayvy The-Wayvy Aug 30, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Proof of Regulatory Capture vs. Proof of Space-Time

Agree with Arthur.

A fair, transparent chain with a lower barrier to entry will attract more SPs.
All else equal that will mean more storage, cheaper storage, and better storage.

The notary system doesn't incentivize useful storage.
It incentivizes bribery and fraud.
This is because it relies on unverifiable computation.
The sophistication of the fraud will scale with the size of the bureaucracy and the number of rules.

I don't want the miners on this network to spend all their time hobnobbing with governance people at FF, making bogus data sets, faking retrieval, bribing notaries, selling DC, etc.
It's a waste of their time and adds no value to the network.

What is the best use of our limited resources?

Fil+ may have brought some benefits to Filecoin.

But any further work on a permissioned Fil+ has lower ROI than researching permissionless solutions.

I have asked several times for an objective measure of usefulness other than willingness to pay and no one has given me one.

The subjective judgement involved in any human notary system will be a source of never-ending conflict.

The time, energy, and money spent mediating these conflicts and iterating on elections, guidelines, etc. would be better spent working on the actual tech itself.

Please remember: Filecoin can exist without the notary system, but the notary system cannot exist without Filecoin.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

  1. FIL+ is not limited to the notary system. There are 4 different data cap distribution programs within FIL+ - one of which automatically gives any github acct that requests it 64GB of datacap per month.
  2. Data storage clients are a valued and critical constituency within the Filecoin network. If your proposal aims to disenfranchise them in favor of SPs, you need to say so explicitly in your proposal.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If we are going to take the current state of the world where SPs have access and influence ( partly because the network spent the initial years focusing on the supply side) and use that as an argument to build the future state, then how do we create a place for clients in this network? If data clients are arguably a negligible constituency, this should be a concern for us and we should find ways to support this constituency.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@raghavrmadya thanks for your contributions here, but this PR review should be fairly strictly focussed on editorial review of the content. It's not the place to ask or answer questions like this. Please do raise them in the discussion thread item, though.

Copy link

@The-Wayvy The-Wayvy Sep 4, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@raghavrmadya
@momack2

It's too early for the use cases that Filecoin Plus is trying to subsidize

I believe the root of our problem is a lack of patience.

We don't even have stablecoins or a DEX on Filecoin.

In fact we don't even have good oracles.

There is still a lot of work to be done on very basic infrastructure which exists on other chains, ex. Ethereum.

If those chains haven't taken users from their Web2 counterparts then it's definitely too early for us to do so.

We do have real users on the FVM and the miners themselves are users because they pay for blockspace.

Let's not
1.) deny the value of these real users.
2.) try to force certain use-cases and users on the network before Filecoin is ready for them.

Just keep building and have a little faith.

Please keep an open-mind to who the next set of real users and use-cases will be.

It probably will not be what you expect: just like few saw Uniswap or Cryptokitties etc coming.

To this day there are no mortgages or treasury bonds being traded on Ethereum, and it still has a larger market cap than the NYSE and many other TradFi giants.

Our goal should be to differentiate ourselves as much as possible from BTC and ETH.


The choices would largely depend on each SP’s comparative cost of obtaining hardware and pledge tokens, combined with their revenue mix between block rewards and storage payments. SPs with clients offering better pay rates may consistently opt for the minimum pledge. This decision could improve their unit economics, particularly when other providers choose to pledge higher amounts to increase their per-TiB rewards. In this landscape, the higher pledge by some becomes an advantage for those opting for a lower pledge.



## Implementation
-Refer to the specific change [here](https://github.com/filecoin-project/builtin-actors/blob/485778aa23f742af1d1aa57a8f6608a0698d8ee7/actors/datacap/src/lib.rs#L161).


## TODO
- Discuss possible deprecation of the Verifreg Actor and Datacap in future updates.

## Copyright
Copyright and related rights waived via [CC0](https://creativecommons.org/publicdomain/zero/1.0/).