papers/zoo-research-coordination/sections/07-implementation.tex

\section{Implementation}

\label{sec:implementation}

\subsection{Protocol Stack}

ZooCoord is implemented as a layered protocol stack on the Zoo Network:

\begin{figure}[H]
\centering
\begin{tikzpicture}[
    node distance=0.5cm,
    layer/.style={rectangle, draw, minimum width=10cm, minimum height=0.8cm, text centered, font=\small},
]
    \node[layer, fill=red!10] (app) {Application Layer: Experiment Dashboard, CLI Tools, IDE Plugins};
    \node[layer, fill=orange!10, below=of app] (coord) {Coordination Layer: Commitment Manager, FEG Scheduler, Verification Market};
    \node[layer, fill=yellow!10, below=of coord] (exec) {Execution Layer: Container Runtime, Provenance Logger, Secure Aggregation};
    \node[layer, fill=green!10, below=of exec] (storage) {Storage Layer: IPFS/Filecoin (Data), Zoo Chain (Commitments, Proofs, Tokens)};
    \node[layer, fill=blue!10, below=of storage] (network) {Network Layer: libp2p (P2P), Zoo Network Consensus};
\end{tikzpicture}
\caption{ZooCoord protocol stack architecture.}
\label{fig:stack}
\end{figure}

\subsection{Key Software Components}

\begin{table}[H]
\centering
\caption{ZooCoord software components and implementation details.}
\label{tab:components}
\begin{tabularx}{\textwidth}{@{}llcX@{}}
\toprule
\textbf{Component} & \textbf{Language} & \textbf{Lines} & \textbf{Description} \\
\midrule
Commitment Manager & Rust & 8,400 & Commitment creation, signing, deviation tracking \\
FEG Engine & Rust & 12,100 & Graph construction, scheduling, execution orchestration \\
Provenance Logger & Rust & 4,200 & Merkle DAG construction, attestation verification \\
Verification Runtime & Go & 6,800 & Sandboxed re-execution, output comparison \\
Market Smart Contracts & Solidity & 3,100 & Bounties, staking, dispute resolution \\
CLI Tools & Python & 5,600 & Researcher-facing experiment management \\
Dashboard & TypeScript & 9,200 & Web interface for experiment monitoring \\
\bottomrule
\end{tabularx}
\end{table}

\subsection{Integration with Existing Tools}

ZooCoord integrates with established research tools through adapters:

\begin{itemize}[leftmargin=*]
    \item \textbf{Jupyter:} A kernel extension automatically captures cell execution provenance and generates FEG nodes from notebook cells.
    \item \textbf{R/RStudio:} An R package wraps common analysis functions with provenance logging and commitment checking.
    \item \textbf{Nextflow/Snakemake:} Bidirectional translators convert existing workflow definitions into FEGs and vice versa.
    \item \textbf{Git:} A Git hook automatically updates the commitment record when analysis code changes, classifying modifications by deviation severity.
\end{itemize}

% ============================================================
% 8. Evaluation
% ============================================================