Improving Graph Compression for EfficientResource-Constrained Graph Analytics

Recent studies have shown the promise of directly processing com-pressed graphs. However, its benefits have been limited by highpeak-memory usage and unbearably long compression time. In thispaper, we introduce Laconic, a novel rule-based graph processingsolution that overcomes the challenges of restricted memory andimpractical compression time faced by existing approaches. La-conic, for the first time, ensures minimal memory overhead duringcompression and significantly reduces graph sizes, thus reducingpeak memory demand during computations. By employing an effi-cient parallel compression algorithm, Laconic achieves a remarkablereduction in compression time. In our experiments, we compare La-conic with state-of-the-art solutions. The results demonstrate thatLaconic outperforms other methods, reducing peak memory con-sumption by an average of 70% during compression and 66% duringcomputation. Additionally, Laconic reduces rule compression timeby an average of 93% compared to traditional rule-based compres-sion, achieving a 2.47xhigher compression ratio, and providing a2.12xperformance speedup.