PPoS / Consensus / DA §
Research §
- Documents (1,2) have been prepared for the probability comparison between chunk sampling and column sampling. The initial impression leans towards continuing with chunk sampling and integrating a batch proof mechanism into it. This way, light clients can efficiently verify multiple chunks at once without bandwidth issues, based on their preferences.
- WIP: Still working through the Cryptarchia v1 protocol specs, we now have pseudocode for most of the important protocols, but still needs more polishing.
- Proposed the “EIP-1559 with moving target” as the current way to go for the blockspace market (more details to come over the next iteration, as it comes now to the point of integrating all markets).
- Proposed a tx fee mechanism for NomosDA market (that will go into further scrutiny, of course, definition of parameters, viability…).
Development §
- PR #970 (Node) - computed DA bandwidth calculations for sampling differences between full columns and single chunks. This will be revisited later once the DA sampling conversations are finished to move the data into the sampling document.
Blendnet §
Research §
- Analysis of communication (and anonymity) failure in the Blendnet: considered broadcasting on linear trees where (“naturally” occurring) communication and adversarial failures are modelled by two different (random) binary variables. Computed probabilities of communication and anonymity failures. The summary of the above work is here while the detailed analysis can be found here.
Development §
Coordination Layer §
Research §
- Benchmarking of IMT in the PoC. There is a linear decrease in proof times switching from smt to IMT but the constants are still high. With a SMT checking a nullifier tkes 650k cycles, with a 64bit path (full IMT) it takes 130k. Which is a linear reduction but still ~1s on a 3090ti per nullifier. We are skeptical of the potential of the batch mechanism since nullifiers are uniformely distributed.
Development §
Zone and Apps §
Research §
Development §
Development §
- PR #974 (Node) - published on-demand docker image for testnet.
- Update testnet with the working metrics and tracing: previous week an issue was identified that prevented testnet nodes to publish metrics and traces, it was related to some services failing because of non existent risc0 dependency. That dependency sometimes failed to be built because of GitHub download rate limit. Logos organization on GitHub needs to update permissions for packages for them to be public, already reached out to infra team. When image is accessible by the testnet host machine, the grafana dashboard with logs, traces and metrics should be available via http://testnet.nomos.tech/.
- PR #975 (Node) - merged the PingPong example (done the week before).
- PR #39 (Overwatch) - passing initial parameter into
ServiceCore::init.
- PR #40 (Overwatch) - POC: State Saving.
- PR #41 (Overwatch) - resolved old branch’s conflicts.
Miscellaneous §
- 2024 recap published.
- NomosDA vs Other DA Solutions post draft [undergoing review].
- Ouroboros family draft [unfinished].