Prior art
l2trace stands on a substantial pile of prior work. This section catalogs the academic and standards-track research we drew from, organized into four tiers of relevance. Each tier page summarizes the papers with the same shape: paper metadata + problem framing + key data model / algorithm details + what l2trace borrows (or deliberately doesn’t).
| Tier | Topic | Papers |
|---|---|---|
| Tier 1 | AFT-based L2 topology discovery (direct prior art) | Breitbart 2004, Lowekamp 2001, Bejerano 2009 |
| Tier 2 | Network verification + belief checking | Anteater, HSA, ATPG, VeriFlow, NetPlumber, NoD |
| Tier 3 | L2 standards + post-STP fabrics | Perlman STP, SPB, All-Path, PortLand, VL2, TRILL |
| Tier 4 | Bitemporal storage foundations | Snodgrass 2000, SQL:2011 |
For citation purposes, the bibliography has the full DOI lookup table plus the BibTeX entries.
For the cross-cutting “what we chose vs alternative architectures” narrative — including the implementable features we sketched but haven’t shipped — see Prior-art synthesis.
Methodology
Section titled “Methodology”Retrieval and synthesis followed a four-tier reading list. For each paper we pulled the full PDF, extracted text, and read the relevant algorithm and evaluation sections directly — not from memory or secondary summaries. PDFs are archived locally; this site links to DOIs and open-access mirrors where available, but doesn’t host the closed-access PDFs themselves.
Citations are formatted [Author Year, §section] so a reader can
verify any claim against the source.
Conventions per paper
Section titled “Conventions per paper”Each paper gets:
- Paper. Authors, title, venue, year, DOI (or stable URL).
- Problem framing. What the paper is trying to solve in its own terms.
- Data model and algorithm. Key definitions + the operational shape of the technique.
- What l2trace borrows. Or doesn’t, and why.
Where two papers disagree (e.g. Breitbart vs Lowekamp on AFT completeness), that’s called out explicitly in the tier page’s cross-cutting observations.