Filecoin Official: Space Race 2 AMA Summary

The second round of Space Race has begun! You can learn more about it here: https://filecoin.io/blog/space-race-ama-3-summary/

The network benefits from having as much storage as possible, so ideally we want to keep it all. It would be great if the percentage reached 100%. However, there is an important principle to keep in mind: the initialization of the mainnet must be well decentralized. Currently, the power distribution is good, but it may get worse in the next two weeks (one miner increases much more than others, or many miners lose a lot of sectors due to failures/terminations during the 14 days).

SR2 tests the network's continued ability to store data and perform upgrades at scale, securely and efficiently. It also seeks to test storage client workflows at a larger scale. Currently, we are in the charging phase, and the goal for miners in SR2 should be to maintain or expand storage and perform any required maintenance. We will announce more SR2 specific goals and incentives soon.

SR2 will store client data as part of Track 2: Slingshot, which will be released in the coming days. This will be optional for miners, but will help test the demand side of the market.

This is possible, but not recommended. We strongly recommend that you use the same setup you use for mainnet.

For bug reports, using Github's issue board is most welcome. Alternatively, you can post your questions in #space-race, #fil-lotus, or #fil-help.

It will remain single threaded on the CPU.

Early next week.

Miners can always upgrade recognized capacity sectors to sectors with verified client transactions.

The SR: Slingshot Track Rules will be published later this week. Useful data can be found or provided by any of the following: clients, miners, or application developers.

Yes, the faucet can also be used to test Filecoin storage clients and value transfer workflows. For example, if you are building an application that uses Filecoin storage retrieval, or a wallet, you will need to use the faucet.

Yes, SR1 is over. SR2 is about maintaining or growing storage (for miners) and testing real data flows (for clients, miners). Yes, the network continues from SR1 and all storage can be maintained. SR2 exists in part because we want to save and use the amazing amount of storage we built in SR1. Without that storage, we would likely not have SR2.

Space Race 2 does not require registration.

Lotus v0.8 adds the ability to upgrade actor implementations in Lotus - this will help the network respond quickly if any issues arise, and make good use of new improvement suggestions from the community. We expect Lotus and the Filecoin network to continue to improve over time, and want to validate and test our ability to iterate quickly.

We encourage miners to take time during the second space race to adjust their setups as needed. We are working on adding more documentation here on how miners can move data between setups so they can use this time to prepare their hardware and architecture for mainnet.

The main miner reward for Space Race 2 is that miners who maintain storage during this period will be able to transfer some of the sectors sealed during Space Race to the mainnet. Therefore, we recommend that you maintain and continue to prove all sectors that you want to be eligible for transition. This answers the questions many miners have about how they can provide staking security in the early days of the network while continuing to reward the amazing amount of data on the network during Space Race.

Space Race 2 has already started (and uses Lotus v0.7.x). There will be optional releases along the way, and possibly one or more network upgrades.

Message packaging is ultimately a resource allocation problem. Block-producing miners may prioritize their own messages, but this does not mean that they have no incentive to include messages from other miners, or that these block-producing miners can include their own messages for free (the base fee is used by all miners). It's just that when a block-producing miner produces a block, they can pay themselves a gas premium close to zero, but other miners' premiums may be higher. Not including other miners' messages can also bring profits, because some other miners will receive the gas premium.

There is nothing special about participating as a new miner. For an overview of the program structure, click here. To learn how to mine, go to docs.filecoin.io.

We are actively working on making Lotus better at calculating and submitting WindowPoSt messages. This also includes minimizing corruption when things go wrong – we don’t want one bad sector to cause the entire partition (or worse, the entire deadline!) to be slashed. These improvements are in progress (some are already in the latest master branch). In addition, the research team is also working on making WindowPoSt more tolerable to real operational risks without sacrificing too much security and quality of service.

Most of the mainnet parameter proposals are already in the codebase. The research team is working to make error penalties more tolerant to real-world operational failures, especially WindowPost failures, and to address some other slack shortcomings. As a reminder, economic structures and parameters will continue to emerge and evolve over time after the network goes live.

We encourage miners to take the time during the second space race to adjust their settings as needed. This is still WIP (should be completed in the next 24 hours), but we recommend the following workflow to safely migrate sectors.

Like any other blockchain project, chain width is a limited resource. There are many reasons why block producing miners only include their own messages when the network is congested. However, they are still incentivized to include other miners' messages (otherwise other miners will get a gas premium), and including their own messages is not free (all participants pay a base fee).