Reference Library
This public selection contains 23 references, with original publication and primary-source links. Three additional papers supplement the earlier 20-entry selection. The working Zotero collection has 103 parent references, including 98 stored paper or book PDFs and five supporting records. The public selection and the private manuscript bibliography have different purposes. Their counts do not measure reading depth or research completion.
Download the current BibTeX selection
The foundational supplement used targeted pages and sections for the explanations in the opening. It is not a fresh full-text appraisal of the entire bank. The private records preserve those limits. The collection-year chart remains a historical 87-record snapshot; it does not measure field-wide publication growth or establish novelty.
November 4 Reading Follow-up
November 4, 2026. The three additions retain the reading boundaries below.
| Source | Place and reading boundary in this project |
|---|---|
| He et al. (2025) | Central multi-router detection, Louvain communities and VAR prediction. Selected Sections 5.2–5.3, 6.1 and 6.3 were read after the earlier abstract-only check. This is not a complete appraisal or reproduction. |
| Nagy, Orosz, Bozsó, & Skopkó (2025) | Acquired paper on cross-domain DDoS datasets. SCLDDOS2024 is the dataset name; the paper is dated 2025. Full appraisal and suitability for the study remain pending. |
| Barrett, Dorai, Li, & Rajaganapathy (2026) | Acquired paper on streaming intrusion detection and drift estimation. Full appraisal and any role in the bounded study remain pending. |
The two additional RFC records support architecture terminology and precedents. They are linked in the provider architecture assessment, outside this selected paper bibliography. Neither acquisition nor citation establishes a novel mechanism or a new result.
Foundational Reading Route
| Source | Place and limit in this project |
|---|---|
| Mirkovic & Reiher (2004) | Historical DDoS terminology and taxonomy, not current prevalence |
| Jordan & Mitchell (2015) | Orientation to the learning task, without selecting a detector |
| LeCun, Bengio, & Hinton (2015) | Neural representations and backpropagation, without claiming DNN superiority |
| Chandola, Banerjee, & Kumar (2009) | Anomaly detection, available labels and score-to-decision roles |
| Rumelhart, Hinton, & Williams (1986) | Early backpropagation reference; scanned equations were not individually checked |
| Bottou, Curtis, & Nocedal (2018) | Gradient training and optimization, with theory assumptions kept explicit |
| Kairouz et al. (2021) | Cross-silo FL and organizational collaboration, without an automatic privacy guarantee |
| Mirsky, Doitshman, Elovici, & Shabtai (2018) | Conditional autoencoder illustration, not the final model or a provider-scale claim |
Existing Domain and Method References
| Source | Place in this project |
|---|---|
| Lavaur (2024) | Main dissertation reference for federated intrusion detection |
| McMahan, Moore, Ramage, Hampson, & Agüera y Arcas (2017) | Federated averaging reference |
| Kingma & Ba (2014) | Adam optimizer; this entry cites the 2014 arXiv record |
| Sellier (2023) | Source of the proposed quantum-inspired training approach |
| Lübben & Pahl (2025) | Decentralized anomaly detection |
| Lübben & Pahl (2024) | Selective federation for device-specific anomaly detection |
| Husák, Manoj, & Kumar (2024) | Operational requirements and limits of machine-learning-based intrusion detection |
| Li, Jiang, Zhang, Kamp, & Dou (2021) | Local batch normalization for federated learning with feature shifts |
| Sun, Esaki, & Ochiai (2021) | Segmented federated intrusion detection in multiple LANs |
These references retain their earlier reading boundaries. The segmented-FL screen covered accessible author prose, with image-only tables unread. The two Lübben and Pahl papers have working Zotero records; project-linked open deposits provide the 2025 article and 2024 conference paper. The newly added foundations do not establish a new exhaustive novelty search or prove that collaboration helps.
Supporting Engineering and Contemporary Evidence
| Source | Place and limit in this project |
|---|---|
| Dobbins et al. (2021) | Informational provider and mitigation use cases, not an FL effectiveness experiment |
| Cloudforce One (2026) | Provider-attributed Q3 2026 operational observations, not global attack prevalence |
| Cloudflare (2026) | Counting-method limits that must accompany use of the provider report |
The industry report and its measurement documentation serve different roles from peer-reviewed research papers. Keeping them in the collection does not change their evidence type. The publication records remain the sources to cite; this website connects them without replacing original authorship or DOI attribution.
Dataset Access
- CICDDoS2019 at the Canadian Institute for Cybersecurity.
- FLNET2023 source repository and dataset links.
These are access points, not exact experiment versions. Dataset papers and actual acquisition versions retain their own working records. This page redistributes no dataset files or publisher PDFs.