10 KiB
| title | draft | lastmod | tags | ||
|---|---|---|---|---|---|
| 2023 September Nomos Monthly Report | false | 2023-09-27 |
|
Executive Summary
The Nomos team continues to stay focused on research related to the fine-grained details of the Base Layer specification and implementation. The addition of a Project Manager is expected to not only expedite the research by allowing the lead to dive deeper into the issues involved but also make that work and progress more transparent.
Mehmet joins the team next month to fulfill a long needed dedicated role in research privacy and zero-knowledge technology as it pertains to the Nomos requirements.
The whitepapers plan to be polished and published by early next month which not only details the current specifications but also known problems that need to be solved.
Key Updates
Personnel
- Filip has joined as a Project Manager (last month) to assist in various activities. His position is to sit in between the Insights team and the Nomos project to facilitate development tracking and resource allocation. It is anticipated that his involvement will speed up development as current resources are then freed to focus on research and development. Until demands within Nomos require full-time engagement from him, he will also be assisting with Vac Program Management.
- After a lot of candidate interviews, Mehmet was offered a position and accepted to focus on the privacy and cryptography research needs within the project. His background in cryptography and security auditing of popular zero-knowledge platforms is expected to be very useful in aiding to architect Nomos. Mehmet's starting date is Oct 2
- Another candidate for this role that was considered, Ramses, has joined the Vac team and will initially be aiding in this work and growing the relationship with their vac/acz/index team. Ramses started in
- An offer has been extended to a candidate for the open role of Applied Network Researcher. He passed all interview rounds but unfortunately passed on the final offer for another opportunity.
- The roles of Applied Network Researcher and Distributed Systems Researcher are still open and candidates are being interviewed.
Milestones
Nomos' roadmap is currently composed of five main sections, each of which is broken up into Research and Development. Obviously, Development lags behind the Research section. These five areas are as follows:
WhitepapersNetwork Privacy and MixnetTestnetPrivate PoSData Availability
The Milestone definitions and timelines are not yet ported to this site, and as such, are private to Logos. They are planned to be publicized on this site by the end of the this month, potentially bleeding into the beginning of October. For those with Notion access, you can view the Milestone definitions here.
The following sections will outline key activity across these milestones. If more information and detail is desired, links can be found under the associated heading among the #Sources and Useful Links Weekly updates section.
Whitepapers
The current Whitepaper drafts are almost ready for publication. They require only "beautification" and a listing of detailed "current unsolved problems" that need to be explored later in the total project. This will be done following the current prioritization of mixnet specification.
This milestone will be completed next month.
Network Privacy and Mixnet
The specification of the mixnet is currently underway and expected to wrap up soon (next month??). This initial specification is modeled after the current State of the Art in the mixnet industry. This is chosen to be critical path based on all the depending architectural decisions that stem from decisions of the networking layer.
Review and analysis of current mixnet literature continues. Throughout this review, a modeling framework was developed in order to help evaluate future (v2) speculative mixnet architectures as compared to the current one. This theoretical framework has already been able to reproduce known results within the industry, such as deriving that the probability of an anonymity failure:
- decreases when increasing the number of layers
- increases when increasing the number of nodes within a single layer.
Framework and analysis details can be found in the overleaf document.
While research explores subtleties in the mixnet specification, development has tackled the foundations of its implementation on top of libp2p within the nomos-node repository. They've logically setup how it connects to the rest of the pieces of the node, setup testing frameworks and node monitoring hooks, and generally got it provisioned for the upcoming testnet.
Testnet
A PoC/Draft Testnet was created and merged via Docker-Compose, then general exploration was done to identify what to measure and how to do it within the nomos-node software.
Simulations of the branch overlay with 1 committee was conducted. Initial results discovered a bug in reproducibility that was fixed. Additional simulations resulted in discovery of inter-module errors with the leader selection. This is currently being explored along with integration of mixnet developments.
Private PoS
Research was conducted in a variety of areas around a Private Proof of Stake spec relevant to the architecture of Nomos. All details can be found in the Whitepaper descriptions.
The initial design was created based on Zcash designs with section details around how "stake" is considered for the Base Layer of Nomos, how "restaking" could work, and various consensus modifications around Carnot. An introduction of "shadows" was done which presents the ideas around the initial voting process. The logic of these "shadows" is currently being fleshed out so that they're more intuitive to understand. One result is that they're now called "Validators" in an effort to keep naming conventions reasonable across the industry.
Much attention was spent on a discovered issue with vote propagation within this construction. This issue is around the implications of vote withholding and how they change as you move up the overlay. The solution being to propagate a map of seen votes upstream alongside the vote, thus increasing transparency of participation.
The concept of "Hastiness" has been introduced to describe the leader's ability to decide whether or not to include votes from the root committee if the threshold is reached before they've had the opportunity to conclude. More research is underway to detail the implications of this decisions with respect to payout, latency, and security.
Most notably, an extension (or more constrained parameterization) of Carnot was initiated and is underway (expected by end of Oct) as a consequence of needing to mitigate a signature aggregation issue. This extension requires 2/3 votes to be collected before a decisions is made.
Data Availability
Research continues on the various available Data Availability schemes and their trade-offs which resulted in the ability to make some decisions on the Nomos specification and identified a more specific personnel requirement for specialized cryptographic expertise. This research demand will now be filled by the new addition to the team and progress is expected to accelerate.
An analysis of the impact of node decay in Data Availability was performed and presented in a Logos Research Seminar. This research resulted in a draft specification of private DA. This then lead to the first draft of a complete privacy solution for the networking layer for consensus and DA.
The architecture of Data Availability services was fleshed out within nomos-node software. Data dissemination implementation was completed and the mempool for certificates is currently in progress.
Perceived Changes in Project Risk
Privacy remains at the forefront of this project's desired requirements. As such, it is important to define what types of privacy the project can offer and then detail exactly how various technology provides it. This is a "ground up" process that starts at the lowest layers and propagates up through the stack. Due to the continued exploration and designing of the Base Layer, it remains to be seen how the currently designs will impact upper layers, namely "Execution Zones." The on-going development of the vac/zkvm/index incubation project within Vac raises the risk of incompatibility between the two as both projects are optimizing for their respective domains in parallel and without much communication.
Future Plans
Insight
It is expected that the entire Nomos roadmap will be specified within this site and the weekly reporting process will be in line with the Milestones defined therein.
A Logos Collaborations section will be included next month to highlight differences in alignment with the Logos Collective as well as cross project collaboration updates.
Depending on the uptake and viability of the Waku reporting process to other projects, then a myriad of quantitative measures will be included in the next monthly report.
Project
Next month focuses on finalizing and publishing:
- the first version of the Whitepapers
- emphasis that specification details will focus on the Base Layer implementation, leaving unknowns and explicit detail of above layers to be explored as open problems.
- a Nomos Testnet Client and Node implementation
- a viability analysis of an embedded mixnet
- a specification and pseudocode for the extended 2/3-vote aggregation Carnot consensus algorithm (Vac dependency: vac/dr/consensus/nomos/carnot-2-3rds-vote-aggregation)
Sources and Useful Links
Weekly updates referenced