2014/12/28 by Amlan Kusum, Kusum, Amlan, Iulian Neamtiu +3
Computer Science · #Cloud Computing and Resource Management #FOS: Computer and information sciences #Programming Languages (cs.PL) #Software Engineering Research #Software System Performance and Reliability
paper · pdf · doi:10.48550/arxiv.1412.8120
openalex publication_date 2014/12/28 · openalex created_date 2025/10/24 · openalex updated_date 2026/07/28
Graph processing is used extensively in areas from social networking mining\nto web indexing. We demonstrate that the performance and dependability of such\napplications critically hinges on the graph data structure used, because a\nfixed, compile-time choice of data structure can lead to poor performance or\napplications unable to complete. To address this problem, we introduce an\napproach that helps programmers transform regular, off-the-shelf graph\napplications into adaptive, more dependable applications where adaptations are\nperformed via runtime selection from alternate data structure representations.\nUsing our approach, applications dynamically adapt to the input graph's\ncharacteristics and changes in available memory so they continue to run when\nfaced with adverse conditions such as low memory. Experiments with graph\nalgorithms on real-world (e.g., Wikipedia metadata, Gnutella topology) and\nsynthetic graph datasets show that our adaptive applications run to completion\nwith lower execution time and/or memory utilization in comparison to their\nnon-adaptive versions.\n