6Block Filecoin Mining — Lee White on WaitSeed and Pledge Collateral
Shared by: Lee White, Core Engineer at 6block The core development engineer
The following article is a summary from the live video sharing by 6 Block’s core engineer Lee White on July 28, 2020, and you can find the video in Chinese at the 6Block Youtube Channel here.
The only thing that has not changed is that it has been changing. This is the best way to describe current stage of Filecoin mining.
— — Lee White
Protocol Labs, the official team behind Filecoin, has been making changes to its economic model, underlying principles, code, etc. We need to adapt to the official changes and understand the reasons for the changes to mine more efficiently. Today, I will share three changes that I think have a relatively significant impact: WaitSeed, Pledge Collateral, and Block reward.
WaitSeed
Speaking of WaitSeed, we must first mention the sealing process. The following picture shows the detailed process of sealing.
This is the sealing process demonstrated by the Protocol Labs. You can see that the first state is “Empty,” and then Packing will fill up the remaining hard disk space. After Packing, there will be unsealed sectors, and then follows the steps of PreCommit1, PreCommit2, PreCommitting, WaitSeed, Committing, and CommitWait, and finally reach FinalizeSector and Proving. Among them, I believe everyone is already very clear about PreCommit1, PreCommit2, and PreCommitting is to submit some results of PreCommit2 on-chain. Today, we will be looking at WaitSeed, which is after PreCommitting, we have to wait for a certain number of blocks to obtain the latest block as a random number, and then extract data from the 11-level sector for the Commit. The primary purpose of WaitSeed is to get random numbers.
Here is one detail I want to point out for WaitSeed:
The rand height is obtained by adding the following two values, PreCommitEpoch refers to the height of the message submitted by PreCommit, and ChallengeDelay refers to the system’s long delay before allowing you to use the latest height as the seed. The initial value of ChallengeDelay was 10, but now it has been changed to 150:150 height * 25 seconds / 60 seconds = 62.5 minutes, which means that miners need to wait 62.5 minutes before executing PreCommit1 and PreCommit2.
For such a waiting mechanism, the explanation given by Protocol Labs is for security considerations and to avoid this value being predicted in advance by miners.
So what impact does WaitSeed have on the sealing process?
Let’s take our F3 series with 256 GB memory as an example. Under normal circumstances, you can run 4 PreCommit1 in parallel at the same time and then execute PreCommit2 sequentially. With WaitSeed, it means that there will be a 62.5 — minute waiting period between PreCommit2 and Commit. If the execution time of 3 PreCommit2 is less than 62.5 minutes, then the machine will be idle for a while.
If you want to solve this problem, you can do another round of PreCommit1+PreCommit2 after completing the four PreCommit2. After all of them are completed, there are eight waiting states. For our currently available Filecoin mining solutions, we provide 8T hard drives; there will be enough space to complete it.
Pledge Collateral
Previously, most of the collaterals are not upfront. When a sector is sealed, the miner does not need to pledge collateral but after sealing and getting reward, the reward will lock up for a while as collateral.
And now Filecoin officially launched the concept of upfront pledge collateral in the latest Calibration version, which means that every sector sealing requires a certain amount of FIL as collateral in advance to start mining.
The formula for calculating the pledged collateral is as follows:
Pledge collateral = 20-day mining output of a sector + 0.3 * The current FIL circulation volume of a sector /the effective computing power of the entire network.
As time goes by, the former declines significantly, while the latter is unknown. The overall situation is a slow decline. Let us take a closer look at the 20-day mining output of one sector and current FIL circulation:
Mining output for 20 days in a sector = 20 *(86,400/25)* sector effective computing power/network effective computing power
Current FIL circulation = 2 billion -the balance of the reward address (t02)-the balance of the market (t05)-the burnt coins (the balance of t099)-the entire network pledge collateral
It is worth mentioning that the pledge collateral coins need to be locked for 180 days and then released within the next 180 days linearly. Under such an economic model, most people will not choose mining but simply trade the coins. Therefore, I personally am not very optimistic about such an economic model. I believe that the Protocol Labs will optimize and modify such an economic model in the future.
Block reward
We know that in the Filecoin testnet, when the storage power exceeds 20%, it will be suppressed. Ideally, the reward should be the same as the percentage of the total network computing power. Here is a mathematical model. The reward obtained by miners at a block height should conform to the following Poisson distribution:
The reward of a miner at a block height = basen, n obeys the Poisson distribution of λ=5, and the expected and variance of the Poisson distribution is λ.
With a mathematical model such as the Poisson distribution, it can be guaranteed that the percentage of computing power is the percentage of the block reward.
The mined FIL tokens will still be locked as collateral. It will be locked for 20 days and then released in 180 days linearly. It is foreseeable that the upfront pledge collateral puls the lock-up will lead to very poor initial liquidity. FIL tokens become scarce, and miners need a large amount of FIL as mining collateral, and therefore create a huge demand to purchase the FIL token and drives the price to go up.
To sum up, the three changes mentioned above will have a greater impact on Filecoin’s economic model. The Protocol Labs is currently seeking feedback from miners, and hopefully, there will be reasonable changes before the mainnet launch.
Okay, that’s it for today’s sharing, thank you!
Reference
1. Sealing process: https://github.com/filecoin-project/storage-fsm/blob/master/fsm.go#L159
2. WaitSeed details: https://github.com/filecoin-project/storage-fsm/blob/master/states_sealing.go#L208
3. Details of WaitSeed changes: https://github.com/filecoin-project/specs-actors/pull/657
4. More Filecoin data: https://filfox.io
