K-P2PLab v3 Documentation

Wiki home

Repository and quick start · Concepts · Documentation policy

This is the task map for v3 within the unified wiki. Project concepts, research design, and publications have separate pages here; this index covers commands, UI procedures, schemas, formulas, and operational limits.

Reading paths

• First deployment: Swarm → Agent capacity → run an experiment → save results.
• Design an experiment: scenario library → scenario reference → protocol options → worked examples.
• Failed or interrupted run: recovery choices → Agent cleanup and capacity → logs and connection diagnosis.
• Interpret and report: metric definitions → bandwidth scope → result images → Reporting principles.
• Change the implementation: architecture → development → the relevant reference below.

Deployment and operation

Guide	Owns
Swarm deployment	Host preparation, CLI/completion, networking, updates, shutdown, storage, and deployment validation
Agent capacity and placement	Startup defaults, per-Agent overrides, proportional scheduling, reservations, and cleanup
Monitoring and diagnosis	Prometheus/Grafana, metric exposition, access/authentication logs, and SSE diagnosis

Experiment workflows

Guide	Owns
Scenario library	Save, validate, edit, and reuse scenario inputs
Execution and recovery	Run, repeat, stop, continue/retry, progress order, and estimated finish times
Dashboard	Live panels, metric cards, keyboard/touch controls, and connection state
Topology viewer	Node/edge meaning, layers, filters, freshness, lifecycle, and sampled graph evidence
Saved results	Browse, export, delete, retain, and back up source records and analysis files
Result images and comparison	Analysis jobs, repetition means, figures, comparisons, PNG/CSV/JSON exports

Reference

Guide	Owns
Scenario schema	YAML fields, phases/jobs, distributions, network conditions, defaults, and limits
Protocol options	PubSub, score, validator, discovery, and Kademlia configuration
HTTP API and event contracts	Endpoints, authentication, SSE framing, log schemas, download headers, and analysis jobs
Experiment metric definitions	Populations, delivery bounds, latency, duplicates, graph metrics, and estimates
Bandwidth measurement	Measurement layer, rates/bytes, protocol attribution, and collection quality

Worked experiments

Guide	Owns
Continuous churn and publishing	Multi-host walkthrough, cumulative arrival schedule, expected population, and evidence
Prysm scoring with three delay cohorts	Synthetic workload, baseline parameters, cohort scheduling, and interpretation limits

Development and compatibility

Guide	Owns
Implementation architecture	Components, communication, networks, isolation, and source map
Development guide	Build, test, validate, deploy development images, and maintain documentation
Performance investigation	Historical churn/SSE evidence, benchmark inputs, and validation limits
v2 reproduction mapping	Execution/network mappings, intentional differences, and historical validation
v2 analysis coverage	Source inventory, analysis/visualization mapping, definitions, and evidence limits

The version comparison manifest supports the v2 audit. Runnable YAML inputs remain in the kpl-v3 source repository’s examples/. Historical results and benchmarks describe their recorded environment; they are not current deployment verification.

---

Korean translation: 한국어.