Filecoin: CC Sector Upgrade Guide and Model Research for Storage Providers.
Original Source: Filecoin Network
TL;DR
For Filecoin storage providers (referred to as "SP"), upgrading existing CC sectors seems to be one of the best opportunities: if verifiable transactions and existing CC sectors are available, SP may consider upgrading these sectors instead of encapsulating new ones to minimize new costs.
Generally, the cost difference between upgrading CC sectors and adding new transaction native sectors is small and should not hinder SP from considering the upgrade strategy.
The results may vary depending on the SP, so we invite you to do your own research.
Background
In this article, you will see:
The different storage channels available in the Filecoin network
The expected cost models using captures and non-captures
In the Filecoin network, storage providers can encapsulate two different types of sectors:
CC sectors: Data is set to zero by default
Transaction sectors: Data is obtained from storage clients
Thanks to FIP 0019 ("SnapDeal") and the introduction of the ReplicaUpdate method, it is now possible to introduce data into CC sectors. This means that storage providers who want to store user data currently have two different options:
Encapsulate transaction sectors (i.e., encapsulate new sectors using user data)
Through upgrading the CC sector on SnapDeal, user data is introduced. We differentiate
Upgrade existing CC sectors
Package new CC sectors and upgrade them immediately (CC sector upgrade pipeline)
Our model aims to compare all different strategies of storage providers in storing user data.
Results
Upgrade Existing CC Sectors vs. Transaction Sectors
According to our model, upgrading existing CC sectors appears to be one of the best opportunities for Filecoin storage providers.
Our model predicts that the cost of upgrading existing CC sectors is equal to or even lower than the cost of adding new transaction sectors in the network (assuming no FIL leasing cost and not accounting for the past packaging cost of existing CC sectors). The cost difference as of July 7, 2023, is $0.05 per 32 GiB sector ($1.6 per TiB). As this is based on dynamic calculations (especially gas fees), the data may vary. Furthermore, future protocol changes (see below) could increase this difference and favor the upgrade of CC sector pipeline.
CC Upgrade Pipeline vs. Transaction Sectors
If access to validated transactions is unavailable, starting with CC sectors and upgrading them afterwards is also feasible.
In fact, the difference between the CC sector upgrade pipeline and the transaction-native sector pipeline is not significant. And in our current model, we have not considered miscellaneous costs such as electricity, bandwidth, and labor costs... Considering that the total cost difference after introducing miscellaneous costs will be negligible, we expect the actual difference to be even closer.
Further observation of validated transaction validity
If there are validated transactions but no existing CC sectors available, storage providers should consider using validated transactions to encapsulate new sectors.
Future scenarios
Possible Changes in the CC Sector Upgrade Pipeline
Cheaper CC sector acquisition
Our model assumes that the cost of encapsulating CC sectors and transaction sectors is the same. If we assume that the cost of acquiring CC sectors is lower than the cost of encapsulating new transaction sectors, then the profit margin of the CC sector pipeline will be close to (or even higher than) the profit margin of the transaction sectors. Factors that may contribute to lower CC sector acquisition costs compared to encapsulating new transaction sectors include:
CC sector-focused SaaS pipeline/improvements
Pre-packaged sector market
Gas cost accounting for transaction and sector (re)activation
One of the prominent differences between CC sector upgrades and transaction sectors is due to gas cost accounting. Specifically, as of June 7, 2023:
Individual sectors with transactions do not need to pay gas fees for sector activation and transaction activation (subsidized by Cron).
Sectors encapsulated through the CC sector pipeline pay gas fees for sector activation and transaction activation.
This is a flaw in the protocol that prevents a fair comparison of the two pipelines. There may be some changes in the future to account for gas fees for rebalancing trade activation. For example, please see this discussion. In this case, we expect the cost of CC sector upgrades and transaction sector pipelines to be closer.
Reducing SnapDeal costs
Another option to improve the CC sector upgrade pipeline is to reduce the validation cost of the SnapDeal protocol. Some people have already begun trying to address this issue (see FIP discussion 645).
Model
In our model, we explore various strategies (i.e. pipelines) for storing data on the Filecoin network. The model uses basic SP resources as inputs and relies on some realistic assumptions to calculate estimated costs and returns for each pipeline.
Assumptions
32 GiB sectors
Transaction size is the same as sector size (i.e. 32 GiB). Please note that this is consistent with the current situation (average verified transaction size is around 30 GiB, see here for relevant statistics).
Only verify transactions, assuming transaction duration matches the remaining sector lifespan at activation (i.e. sectors can be extended if needed).
Definition
For more details, please refer to this webpage.
