added nomos weeklies

This commit is contained in:
Corey 2023-09-26 15:35:22 -04:00
parent 9f1d58f257
commit c7c32c6511
3 changed files with 146 additions and 1 deletions

View File

@ -25,4 +25,5 @@ lastmod: 2023-09-13
Weekly updates referenced
- [[nomos/updates/2023-09-04|2023-09-04]]
-
- [[nomos/updates/2023-09-19|2023-09-19]]
- [[nomos/updates/2023-09-26|2023-09-26]]

View File

@ -0,0 +1,77 @@
---
title: 2023-09-19 Nomos weekly
tags:
- nomos-updates
date: 2023-09-19
lastmod: 2023-09-26
draft: false
description: 2023-09-19 Weekly update of Nomos
---
## `nomos:`
---
### `network privacy and mixnet:`
#### `research:`
- Review of the “Untraceable electronic mail, return addresses, and digital pseudonyms” paper. Review of the “The Loopix anonymity system” paper.
- Notes provided in the overleaf document https://www.overleaf.com/read/rybwvjftfrrg
#### `development:`
- Polishing mixnet to make it ready for the testnet:
- Helping preparing mixnet deployments: https://github.com/logos-co/nomos-node/pull/408
- Reviewed mixnet connection management
- Refactor the libp2p network backend: https://github.com/logos-co/nomos-node/tree/libp2p-refactor-yjlee and based on that prepare to measure the quality of unobservability - https://github.com/logos-co/nomos-node/issues/391
- Block building: https://github.com/logos-co/nomos-node/pull/401
- Fountain codes: https://github.com/logos-co/nomos-node/pull/407
- Got the mixnet retry PR merged: https://github.com/logos-co/nomos-node/pull/386
- Finish the concrete error implementation for mixnet and ready for merge: https://github.com/logos-co/nomos-node/pull/405
- WIP: fan-in message passing model for mixnode: https://github.com/logos-co/nomos-node/tree/mixnet-fan-in
### `testnet:`
#### `development:`
- Testnet preparation: https://github.com/logos-co/nomos-node/pull/410
- Testnet POC with libp2p merged: https://github.com/logos-co/nomos-node/pull/364
- The POC for testnet using etcd and docker compose was reviewed and merged. The mixnet functionality will be added on top of this.
- Simulation branch overlay with 1 committee fix: https://github.com/logos-co/nomos-node/pull/402
- A bug where branch overlay results different from tree overlay with one committee blocked research team to use simulations results. Fixed it.
- Improvements in CI:
- https://github.com/logos-co/nomos-node/pull/409
- https://github.com/logos-co/nomos-node/pull/411
- https://github.com/logos-co/nomos-node/pull/412
- Minor improvements to remove annoying red x's from our CI
### `private PoS:`
#### `research:`
- **Shadows logic:** Looking at how to describe the logic of the shadow in the most clear way: It will be divided into a set of modules, each module taking care of processing a separate communication channel.
- All channels have their logic described in adequate modules and with references to self-descripting functions. However, some of them (like how exactly to aggregate votes) must yet to be defined.
- **Hastiness issues:** In short, the leader, in order to limit the cost of vote aggregation can decide to not to include votes from top committees (and root in particular). This is an acceptable strategy and will lead to a correctly formed aggregation proof. The proof will include a global threshold of votes from lower committees but not from the top committees (and root committee in particular). The impact of this leader's hastiness does not break the security of the protocol as a threshold of votes is correctly gathered. However, it may limit rewards from the top committees (and root in particular), as the votes from those committees may not be needed to reach the threshold. More on that under the issues section of the PPoS doc.
### `data availability:`
#### `research:`
- First stab at privacy solution for the network layer for consensus and DA: https://www.notion.so/Practical-Private-Addressing-Network-Privacy-Component-2-2b9b4923124a4fdb81dba5d2bba1d289?d=99166164267a46589c5715175e1b3657#5e27d2010d30468f9d8f0d0928b9c639
- Init survey of SoA in network privacy alternative solutions
#### `development:`
- DA nomos core kickstart, added different pieces that were missing for abstractions: https://github.com/logos-co/nomos-node/pull/390
- Added attestation trait
- Added certificate trait
- Added DaProtocol trait that abstracts encoding/decoding, and put the pieces together for blob+attestation+certificate handling.
- Refactored (moved and restructured) da modules
- Refactor and improve common traits - https://github.com/logos-co/nomos-node/pull/395
- Implement a simple da protocol with full replication - https://github.com/logos-co/nomos-node/pull/400
- Implement a command to disseminate blobs through the network - https://github.com/logos-co/nomos-node/pull/396
- Added da-service to nomos node - https://github.com/logos-co/nomos-node/pull/404
- Housekeeping:
- https://github.com/logos-co/nomos-node/pull/403
- https://github.com/logos-co/nomos-node/pull/388
- https://github.com/logos-co/nomos-node/pull/399

View File

@ -0,0 +1,67 @@
---
title: 2023-09-26 Nomos weekly
tags:
- nomos-updates
date: 2023-09-26
lastmod: 2023-09-26
draft: false
description: 2023-09-26 Weekly update of Nomos
---
## `nomos:`
### `network privacy and mixnet:`
#### `research:`
- With the assumption that nodes of a mixnet are selected without replacement, we performed analysis with Byzantine node presence (for specific widths and lengths). This gives the probability that there is at least one path where all nodes are Byzantine (“anonymity” failure)
- Evaluation in the “Loopix” paper used the mixnet with n1=2 and L=5
- Conclusion: probability of anonymity failure decreases when increasing length and increases with increasing width
- Notes provided in the overleaf document https://www.overleaf.com/read/rybwvjftfrrg
- Discussions on how to model the network privacy for analysis; viability of the embedded mixnet design
- Notes (WIP): https://www.notion.so/Network-Privacy-2dabf0aa878744e299b2ebb97120876f
#### `development:`
- Making integration tests work with FlatOverlay and RandomBeacon:
- https://github.com/logos-co/nomos-node/pull/425
- https://github.com/logos-co/nomos-node/pull/426
- Some integration tests are randomly failed. Debugging them: https://github.com/logos-co/nomos-node/pull/437
- Refactoring libp2p network layer: https://github.com/logos-co/nomos-node/pull/417
- Add missing error handlings in mixnet: https://github.com/logos-co/nomos-node/pull/436 (+ more coming soon)
- Trying to enable gathering metrics for libp2p (but needs to be discussed about how this can be used with our existing metrics service): https://github.com/logos-co/nomos-node/pull/431
- New mixnet message handle PR: https://github.com/logos-co/nomos-node/pull/435
- QC checks: https://github.com/logos-co/nomos-node/pull/430
### `testnet:`
#### `development:`
- TreeOverlay in Nomos node:
- https://github.com/logos-co/nomos-node/pull/415
- https://github.com/logos-co/nomos-node/pull/423
- After adding tree overlay to Nomos node, integration tests started failing. Main reason was that the leader didn't spawn in time and nodes failed to send their votes. This mainly affected the happy path test. Work will be merged once the issues are fixed.
- Testnet with mixnode: WIP
- Work on the mixnet node in testnet continues. Ongoing inter-team discussions in regards to how we should monitor and extract info from the network. The PR for libp2p metrics might be the way to go.
- CI chores: https://github.com/logos-co/nomos-node/pull/432
### `private PoS:`
#### `research:`
- Add missing function descriptions and finalize structure definitions.
- Defined/redefined structures used in all algorithms that required a rewrite.
- Updated the terminology and made the Shadows name obsolete, and now they are called Validators (for synchronization with other PoS designs).
- The validators logic was redesigned, improved and updated accordingly.
- The same was with the ledger/transactions part, and now they form a complete logic.
- Readability: the specification part was updated. The rest of the document is out of sync and needs to be revised as the focus was put on the specification.
- Critical analysis: an issue was identified and described (under the issues section) that touches on a problem where insufficient number of votes are aggregated and the need for an additional voting round before commencing an overlay tree reconstruction. This can be mitigated with an additional votes collection from late voters (before the tree timeout) or increased level of votes that are collected during initial voting collection.
- Review of the Delegation and Validation rewards document by Frederico.
### `data availability:`
#### `:development`
- Dissemination ready
- https://github.com/logos-co/nomos-node/pull/416
- https://github.com/logos-co/nomos-node/pull/400
- https://github.com/logos-co/nomos-node/pull/404
- Mempool for certificates in progress