Another testnet is under way, we’re keeping an eye to see if our recent improvements are improving stability. Do join in if you have time. Nodes still have to be run from the cloud until
libp2p has better QUIC support, but clients can run from anywhere.
We’re hoping to see improved node stability, but we’re still seeing some memory issues. Looking at dead nodes’ logs we tend to see a flurry of messaging just before they keel over, which is something we’re investigating now.
Some of you have noticed a bustle in the DBC hedgerow on GitHub. Indeed, as we mentioned last week this is very much the focus of our attention now, and things are ticking along nicely.
We have the faucet up and running so testnet DBC allocation will be self-service, and we can ensure people have to pay for chunk uploads, with payment for registers pretty much ready too. Then it’s on to calculating how much storage is worth for nodes, and how to reuse network DBCs, but we’re a good way along implementing a basic version of paying for data. After that it will be mechanisms for price discovery, for which we also have a few options to try.
On DBCs @bochaco has implemented a change to make payment for uploads obligatory feat: make Chunks storage payment required by bochaco · Pull Request #533 · maidsafe/safe_network · GitHub. He is looking at options to make the spent proofs smaller. Rust crates are a bit basic here, so we may have to wait or roll our own.
And DBCs will now be stored in kad
record_store - the last of our data types to go in @roland’s been helping here too.
@Qi_ma is looking at a Windows-specific issue with this, where storing DBCs in record store is causing CI to fail. He’s also looking at making the self encryption crate configurable, so that through testing we can optimise the chunk size.
@anselme is pushing the faucet into the local testnet binary so we can give it a thorough check before moving it to the testnet tool proper.
@aed900 is looking at connections timing out and how we can notify users when that happens. He’s also been delving into the QUIC support situation with
@bzee is looking at more libp2p improvements and evaluating the connections fix we spoke about last seek. The current testnet should be a help there.
@chris is looking to refactor the testnet tool (the program we use to launch testnets on Digital Ocean and AWS) to remove some complexity that’s built up.
Finally, @joshuef is investigating ongoing memory and processing issues, trying to identify any potential bottlenecks that might be causing message backlogs.
Feel free to reply below with links to translations of this dev update and moderators will add them here:
As an open source project, we’re always looking for feedback, comments and community contributions - so don’t be shy, join in and let’s create the Safe Network together!