{"schema_version":"0.1","map_id":"paper-66-map","publication_id":66,"publication_anchor":"paper-66","slug":"paper-66","canonical_path":"/knowledge/papers/paper-66/","machine_path":"/knowledge/papers/paper-66.json","root_node_id":"paper-66","stage":"mapped_draft","contribution_type_vocabulary_version":"0.1","contribution_types":["algorithm"],"title":"Traffic Analysis by Adversaries with Partial Visibility","short_title":"Traffic Analysis with Partial Visibility","year":2023,"venue":"28th European Symposium on Research in Computer Security (ESORICS 2023)","publication_status":"Published · proceedings issued in 2024","topic":"privacy-identity","labels":["Theory","Applied"],"ai_ml_labels":["AI for security"],"authors":["Iness Ben Guirat","Claudia Díaz","Karim Eldefrawy","Hadas Zeilberger"],"keywords":["traffic analysis","mix networks","partial visibility","Bayesian inference","Metropolis-Hastings sampling","anonymity measurement"],"research_question":"How can one estimate what a traffic-analysis adversary can infer when the adversary sees or compromises only selected parts of a mix network and the remaining message-flow trace is hidden?","central_answer":"The paper models visible and hidden flow matrices, factors an adversary into goal, prior knowledge, and observation or compromise capabilities, and uses Metropolis-Hastings sampling to approximate a posterior distribution over hidden traces. A nine-mix case study demonstrates how different partial-visibility adversaries can be represented and sampled; it is a framework and proof-of-concept evaluation, not a universal de-anonymization rate.","curation":{"drafted_at":"2026-07-11","drafted_by":[{"actor_type":"ai","name":"OpenAI Codex","role":"full-text algorithm extraction, evidence mapping, and initial assessment"}],"method":"Source-grounded review of the complete 20-page author-hosted manuscript, including visual inspection of the title and evaluation pages, cross-checked against the official DOI.","source_scope":"full_source_audit","approval":{"status":"pending","note":"AI-authored source map awaiting author verification; adversary-model interpretation, evaluation readings, and ratings remain provisional."}},"sources":[{"id":"source-paper-66-local-pdf","type":"author_hosted_copy","title":"Traffic Analysis by Adversaries with Partial Visibility","url":"/pubs/2023/traffic-analysis-partial-visibility-esorics2023.pdf","provenance_category":"author","retrieved_from":"https://www.esat.kuleuven.be/cosic/publications/article-3637.pdf","media_type":"application/pdf","sha256":"ebc2f1971bee7b977274b374f76302ad41e888a6be507d85fed1eddcab348c0a","page_count":20},{"id":"source-paper-66-official","type":"official_publication_record","title":"ESORICS 2023 publication record","url":"https://doi.org/10.1007/978-3-031-51476-0_17","provenance_category":"official"},{"id":"source-paper-66-citations","type":"citation_index_snapshot","title":"OpenAlex record W4390698000","url":"https://api.openalex.org/works/W4390698000","accessed_at":"2026-07-11"}],"source_anchors":[{"id":"anchor-paper-66-problem","source_id":"source-paper-66-local-pdf","label":"Partial-visibility problem and contributions","locator":"Abstract and Sections 1-2, PDF pages 1-4","url":"/pubs/2023/traffic-analysis-partial-visibility-esorics2023.pdf#page=1"},{"id":"anchor-paper-66-adversary","source_id":"source-paper-66-local-pdf","label":"Adversary goals, prior knowledge, and capabilities","locator":"Section 3, PDF pages 4-6","url":"/pubs/2023/traffic-analysis-partial-visibility-esorics2023.pdf#page=4"},{"id":"anchor-paper-66-matrices","source_id":"source-paper-66-local-pdf","label":"Matrix trace, observations, hidden state, and constraints","locator":"Section 4, PDF pages 6-9","url":"/pubs/2023/traffic-analysis-partial-visibility-esorics2023.pdf#page=6"},{"id":"anchor-paper-66-sampler","source_id":"source-paper-66-local-pdf","label":"Metropolis-Hastings inference procedure","locator":"Section 5, PDF pages 9-12","url":"/pubs/2023/traffic-analysis-partial-visibility-esorics2023.pdf#page=9"},{"id":"anchor-paper-66-case-study","source_id":"source-paper-66-local-pdf","label":"Nine-mix demonstrator and adversary instantiations","locator":"Section 6, PDF pages 12-13","url":"/pubs/2023/traffic-analysis-partial-visibility-esorics2023.pdf#page=12"},{"id":"anchor-paper-66-evaluation","source_id":"source-paper-66-local-pdf","label":"Implementation, sampling experiments, and convergence checks","locator":"Section 7, PDF pages 13-17","url":"/pubs/2023/traffic-analysis-partial-visibility-esorics2023.pdf#page=13"},{"id":"anchor-paper-66-limitations","source_id":"source-paper-66-local-pdf","label":"Interpretation, limitations, and future work","locator":"Sections 8-9, PDF pages 17-18","url":"/pubs/2023/traffic-analysis-partial-visibility-esorics2023.pdf#page=17"},{"id":"anchor-paper-66-publication","source_id":"source-paper-66-official","label":"Official publication identity","locator":"ESORICS 2023 proceedings, DOI 10.1007/978-3-031-51476-0_17","url":"https://doi.org/10.1007/978-3-031-51476-0_17"},{"id":"anchor-paper-66-citation-count","source_id":"source-paper-66-citations","label":"Dated citation-count snapshot","locator":"OpenAlex cited_by_count was 0 when accessed 2026-07-11","url":"https://api.openalex.org/works/W4390698000"}],"nodes":[{"id":"paper-66","kind":"paper","parent_id":null,"order":1,"epistemic_status":"published","title":"Traffic analysis with partial visibility","summary":"A general inference framework for representing and sampling the hidden parts of a mix-network trace under adversaries with heterogeneous monitoring and compromise capabilities.","source_anchor_ids":["anchor-paper-66-problem"]},{"id":"paper-66-question","kind":"question","parent_id":"paper-66","order":1,"epistemic_status":"research_question","title":"Research question","summary":"What posterior knowledge about sender-receiver flows is available when an adversary sees only a subset of the network's traffic matrices or internal mappings?","source_anchor_ids":["anchor-paper-66-problem","anchor-paper-66-adversary"]},{"id":"paper-66-answer","kind":"contribution","parent_id":"paper-66","order":2,"epistemic_status":"source_asserted","title":"Central answer","summary":"Represent the known and unknown portions of the network as constrained matrices and sample hidden traces from a posterior that combines observations with an adversarial prior.","source_anchor_ids":["anchor-paper-66-matrices","anchor-paper-66-sampler"]},{"id":"paper-66-model","kind":"model","parent_id":"paper-66","order":3,"epistemic_status":"defined","title":"Mix-network and adversary model","summary":"The communication trace is a sequence of matrices over flows between entities and mixes; an adversary is characterized by its inference goal, prior knowledge, and capability to monitor links or compromise mixes.","source_anchor_ids":["anchor-paper-66-adversary","anchor-paper-66-matrices"]},{"id":"paper-66-model-observation","kind":"definition","parent_id":"paper-66-model","order":1,"epistemic_status":"formalized","title":"Observation and hidden state","summary":"The trace is partitioned into an observed component O and hidden state HS. Network conservation, observed counts, topology, and compromised mappings constrain the set of feasible hidden matrices.","source_anchor_ids":["anchor-paper-66-matrices"]},{"id":"paper-66-model-capabilities","kind":"threat_model","parent_id":"paper-66-model","order":2,"epistemic_status":"defined","title":"Monitoring and compromise capabilities","summary":"The framework distinguishes observation of inter-entity traffic matrices from compromise of a mix, which can expose the mix's input-output correspondence under the modeled honest-but-curious behavior.","source_anchor_ids":["anchor-paper-66-adversary","anchor-paper-66-matrices"]},{"id":"paper-66-algorithm","kind":"algorithm","parent_id":"paper-66","order":4,"epistemic_status":"specified","title":"Posterior inference by Metropolis-Hastings","summary":"A Markov-chain sampler explores feasible hidden traces and estimates adversary-relevant probabilities without enumerating the complete hidden-state space.","source_anchor_ids":["anchor-paper-66-sampler"]},{"id":"paper-66-algorithm-transition","kind":"algorithm_step","parent_id":"paper-66-algorithm","order":1,"epistemic_status":"specified","title":"Feasibility-preserving proposal","summary":"The transition chooses a hidden matrix and exchanges entries across two columns in a way designed to preserve the trace constraints and produce another feasible candidate state.","source_anchor_ids":["anchor-paper-66-sampler"]},{"id":"paper-66-algorithm-posterior","kind":"algorithm_step","parent_id":"paper-66-algorithm","order":2,"epistemic_status":"specified","title":"Bayesian acceptance and estimation","summary":"Candidate transitions are accepted according to the Metropolis-Hastings ratio induced by the adversary's prior and observations, after which samples estimate the posterior quantity tied to the attack goal.","source_anchor_ids":["anchor-paper-66-sampler"]},{"id":"paper-66-claims","kind":"claim_group","parent_id":"paper-66","order":5,"epistemic_status":"source_asserted","title":"Principal claims","summary":"The authors claim a flexible representation for partial visibility and a sampling method that can instantiate substantially different traffic-analysis adversaries within one framework.","source_anchor_ids":["anchor-paper-66-problem","anchor-paper-66-evaluation"]},{"id":"paper-66-claim-generality","kind":"claim","parent_id":"paper-66-claims","order":1,"epistemic_status":"demonstrated","title":"Common model for different adversaries","summary":"The case study instantiates both link-monitoring and mix-compromise views, including a global passive adversary and an adversary compromising entry and exit mixes.","source_anchor_ids":["anchor-paper-66-case-study","anchor-paper-66-evaluation"]},{"id":"paper-66-claim-feasibility","kind":"claim","parent_id":"paper-66-claims","order":2,"epistemic_status":"experimentally_supported","title":"Sampling is feasible in the demonstrator","summary":"The Python implementation generates posterior samples for the evaluated nine-mix, three-layer source-routed topology, and the paper reports convergence diagnostics across generated traces.","source_anchor_ids":["anchor-paper-66-case-study","anchor-paper-66-evaluation"]},{"id":"paper-66-evidence","kind":"evidence_group","parent_id":"paper-66","order":6,"epistemic_status":"documented","title":"Evidence and checks","summary":"The paper gives the mathematical representation and sampler, a 2,363-line Python implementation, multiple adversary configurations, repeated traces, and Wilson-score intervals used to assess sampling convergence.","source_anchor_ids":["anchor-paper-66-sampler","anchor-paper-66-evaluation"]},{"id":"paper-66-boundaries","kind":"limitation_group","parent_id":"paper-66","order":7,"epistemic_status":"explicitly_bounded","title":"Scope and limitations","summary":"The evaluation demonstrates the framework on one mix-network model and does not establish universal anonymity loss, production scalability, or resistance to active manipulation.","source_anchor_ids":["anchor-paper-66-evaluation","anchor-paper-66-limitations"]},{"id":"paper-66-boundary-adversary","kind":"limitation","parent_id":"paper-66-boundaries","order":1,"epistemic_status":"model_bounded","title":"Compromise is not an active-attack model","summary":"The compromised mixes reveal internal mappings under the paper's modeled behavior; arbitrary packet dropping, modification, delay attacks, and adaptive malicious control require additional modeling.","source_anchor_ids":["anchor-paper-66-adversary","anchor-paper-66-limitations"]},{"id":"paper-66-boundary-generalization","kind":"limitation","parent_id":"paper-66-boundaries","order":2,"epistemic_status":"benchmark_bounded","title":"Topology and prior dependence","summary":"Posterior conclusions and convergence depend on the nine-mix topology, source routing, traffic assumptions, prior, capability placement, and attack goal; broader empirical coverage is left for future work.","source_anchor_ids":["anchor-paper-66-case-study","anchor-paper-66-limitations"]},{"id":"paper-66-boundary-computation","kind":"limitation","parent_id":"paper-66-boundaries","order":3,"epistemic_status":"scalability_open","title":"Sampling cost remains a boundary","summary":"Metropolis-Hastings avoids exhaustive enumeration but may require substantial burn-in and sampling for large hidden spaces; the paper does not provide a general polynomial-time convergence guarantee.","source_anchor_ids":["anchor-paper-66-sampler","anchor-paper-66-evaluation","anchor-paper-66-limitations"]},{"id":"paper-66-resources","kind":"artifact_group","parent_id":"paper-66","order":8,"epistemic_status":"source_available","title":"Auditable resources","summary":"The complete author copy is checked into this site with page count and SHA-256, and the official DOI establishes the proceedings identity. The paper describes implementation code but this audit did not locate a public repository.","source_anchor_ids":["anchor-paper-66-problem","anchor-paper-66-evaluation","anchor-paper-66-publication"]},{"id":"paper-66-scrutiny","kind":"scrutiny","parent_id":"paper-66","order":9,"epistemic_status":"venue_reviewed","title":"External scrutiny","summary":"ESORICS publication provides venue-level review. No public review reports, independent reproduction, correction, or comparative re-evaluation was located.","source_anchor_ids":["anchor-paper-66-publication"]},{"id":"paper-66-lineage","kind":"lineage","parent_id":"paper-66","order":10,"epistemic_status":"documented","title":"Methodological lineage","summary":"The work connects classical mix-network traffic analysis with Bayesian latent-state inference and Markov-chain sampling, explicitly broadening analysis beyond all-seeing global observers.","source_anchor_ids":["anchor-paper-66-problem","anchor-paper-66-sampler"]}],"relations":[{"id":"paper-66-rel-answer-question","type":"addresses","from_id":"paper-66-answer","to_id":"paper-66-question"},{"id":"paper-66-rel-model-answer","type":"scopes","from_id":"paper-66-model","to_id":"paper-66-answer"},{"id":"paper-66-rel-observation-model","type":"component_of","from_id":"paper-66-model-observation","to_id":"paper-66-model"},{"id":"paper-66-rel-algorithm-answer","type":"realizes","from_id":"paper-66-algorithm","to_id":"paper-66-answer"},{"id":"paper-66-rel-transition-algorithm","type":"component_of","from_id":"paper-66-algorithm-transition","to_id":"paper-66-algorithm"},{"id":"paper-66-rel-posterior-algorithm","type":"component_of","from_id":"paper-66-algorithm-posterior","to_id":"paper-66-algorithm"},{"id":"paper-66-rel-evidence-generality","type":"supports","from_id":"paper-66-evidence","to_id":"paper-66-claim-generality"},{"id":"paper-66-rel-evidence-feasibility","type":"supports","from_id":"paper-66-evidence","to_id":"paper-66-claim-feasibility"},{"id":"paper-66-rel-boundaries-claims","type":"qualifies","from_id":"paper-66-boundaries","to_id":"paper-66-claims"},{"id":"paper-66-rel-resources-evidence","type":"enables_audit_of","from_id":"paper-66-resources","to_id":"paper-66-evidence"},{"id":"paper-66-rel-lineage-paper","type":"contextualizes","from_id":"paper-66-lineage","to_id":"paper-66"}],"assessment":{"id":"paper-66-assessment-2026-07-11","rubric_version":"0.2","assessed_at":"2026-07-11","status":"ai_draft_author_review_pending","note":"These dimensions describe documented support and process, not truth, correctness, or a universal ranking. No composite score is calculated.","axes":[{"id":"epistemic_evidence","level":"high","rationale":"The full source defines the adversary and latent-state model, specifies the sampler, documents implementation and experiments, and reports convergence checks. Evidence remains bounded to the evaluated topology and assumptions.","basis_source_anchor_ids":["anchor-paper-66-adversary","anchor-paper-66-sampler","anchor-paper-66-evaluation"]},{"id":"auditability","level":"high","rationale":"A complete checked-in author copy with fixity and page count and an official DOI make the assumptions, algorithm, and reported evaluation directly inspectable.","basis_source_anchor_ids":["anchor-paper-66-problem","anchor-paper-66-publication"]},{"id":"production_provenance","level":"medium","rationale":"Named authorship, venue, DOI, and an author-hosted manuscript are documented; contributor roles, version history, source-code identity, and tool-use provenance are not.","basis_source_anchor_ids":["anchor-paper-66-problem","anchor-paper-66-publication"]},{"id":"external_scrutiny","level":"medium","rationale":"ESORICS publication establishes venue scrutiny, but public reports, independent reproduction, correction history, and adversarial re-analysis were not located.","basis_source_anchor_ids":["anchor-paper-66-publication"]},{"id":"reception","level":"low","rationale":"The dated OpenAlex snapshot located 0 citations. Under the author-defined rule, 0 through 8 located citations is Low; a zero in one index is not evidence of no readership.","basis_source_anchor_ids":["anchor-paper-66-citation-count"]},{"id":"contribution_significance","level":"medium","rationale":"The paper supplies a reusable framework for partial-visibility traffic analysis and demonstrates it concretely, while broad applicability and independent uptake remain to be established.","basis_source_anchor_ids":["anchor-paper-66-problem","anchor-paper-66-evaluation"]}]},"reception_snapshot":{"as_of":"2026-07-11","method":"OpenAlex work lookup by DOI 10.1007/978-3-031-51476-0_17","citation_count":0,"source_url":"https://api.openalex.org/works/W4390698000","signals":["OpenAlex reported no citing works for the matched record at the snapshot date."],"limitation":"Citation counts are index- and date-dependent; zero in one index does not prove no reception, and counts do not establish correctness."}}