We have updated the Cryptarchia specification with the update epoch stabilization schedule - PR.
Analysis of adversarial inference of relative stake: derived an equation that can be used to infer the Lagrange parameter in the maximum likelihood inference of relative stake - the details of the analysis.
development
Removed assumptions on Carnot being the consensus algorithm in the mempool: PR.
Separate ledger and consensus to prepare for integration: PR.
mixnet (network privacy)
Research
Work in Progress (WIP): The Mixnet Incentivization document has been initiated. Current open questions will be addressed, covering system design, mathematical analysis, and more.
Continuing work on the mixnet with staking, incorporating modifications based on feedback. A section has been added concerning the hidden bonus of deanonymization. It’s a straightforward observation that random assignment of adversarial nodes can, in some cases, lead to a higher number of adversarial paths than naively expected. Also, a section about discussing the latency and anonymity relationship has been started. WIP document.
Analysis of the fraction of compromised paths in the mix network: using an asymptotic lower bound to estimate the probability that the fraction of compromised paths, α, belongs to the interval [α0,α1]. The analysis suggests that in the mixnet of size n=240 with L=3 layers sampled from N=800 nodes, where M=200 nodes are adversarial, α can be almost three times larger than the average (M/N)L which assumes a mixnet of infinite width. The parameters n=240, L=3, and N=800 are currently used in Nym’s mixnet. Here for ¼ of adversarial nodes, the fraction of compromised paths can be as high as 0.05. To compare, the average here is 0.02. Summary is provided here.
development
All PRs for Mixnet v1 implementation have been merged. We’ve taken some additional time to polish the code according to feedback.
One remaining PR we are working on is adding a compilation option to enable mixnet. We’re going to always enable mixnet in production, but we’ve discussed that it’s also good to remain in the libp2p-only compilation mode for development to unblock other dev topics until everything of mixnet becomes stable. Will be finished this early this week.
data availability
research
No current updates.
development
Finished RS core encode/decode: there was an issue with different FFT calls from different libraries that didn’t work and took a while to debug. They use floating numbers, and when rounding or using a big set of operations, precision leads to errors - PR.
DA API mempool tests using a mock implementation PR - The previous PR defined an abstraction for verifying and filtering what to include in the generic mempool; this PR provides mock implementations for TX and Cert verification/filtering. WIP: DA API indexing for data blobs - adding an index to data blob in DA node when the certificate is observed in the block.