Keyboard shortcuts

Press or to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

VLDB 2027 Systems Paper (In Preparation)

Target: PVLDB Vol. 20, Industry Track

Working Title: Samyama: A Unified Graph-Vector Database with Cost-Based Query Planning and Late Materialization

Status: In preparation. Extends Paper 1 (arxiv:2603.08036) with deeper evaluation and VLDB-specific focus.

Strategy

The arxiv preprint covers everything broadly. For VLDB we focus and deepen on three core database contributions:

  1. Late materialization for property graphs (NodeRef / EdgeRef) — 4× traversal speedup, 60% memory reduction
  2. Graph-native cost-based planner — triple-level statistics, plan enumeration, direction reversal, ExpandInto
  3. Hybrid CSR adjacency — frozen tier + write buffer, two-phase bulk loading, billion-edge ingestion at single-server scale

De-emphasize: GPU acceleration, agentic enrichment, optimization solvers, RDF (mention as system features, don’t evaluate deeply).

Add: LDBC SNB comparison, billion-edge evaluation, cross-KG federation at scale.

Key New Results (since arxiv v2)

ResultImpact
11-KG mega-benchmark, 138M nodes / 1.22B edges, 500/500 pass on r7i.16xlargeScale + correctness credibility
Adjacency-aware aggregation (ADR-017)Algorithmic contribution — closes degree-aggregation timeouts on PubMed-scale data
WCO TrieJoin for cyclic patternsAlgorithmic contribution
Edge-arena removal (DS-07c) — COW edges via EdgeViewEngineering contribution — substantial memory savings at 1B-edge scale
MVCC: Snapshot Isolation + Read Committed with write-write conflict detectionConcurrency story
Rayon parallel algorithms (PageRank, LCC, CDLP)Parallelism

Manuscript sources, planning notes, and formatting TODO live in the private samyama-research repository under papers/paper-vldb-2027/.