Query evaluation Depth-first traverse the index tree Bufter management ( LRU Why buffering? Decompression Time Dominates Decompression avoidance
26 Query Evaluation ▪ Depth-first traverse the index tree ▪ Buffer Management (LRU) • Why buffering? Decompression Time Dominates • Decompression avoidance
O utline ■ Introduction XQzip Indexing Data Compression Query evaluation Performance evaluation Conclusion
27 Outline ▪ Introduction ▪ XQzip • Indexing • Data Compression • Query Evaluation • Performance Evaluation ▪ Conclusion
Effectiveness of the sit NO ode Node Data Source Load time Reduction Selection Acceleration XMark 1.64 0.67s 2.15 OMIM 0.240% 0.07s 2. DB BLP 0.049% 1.62S 2. SwissProt 28.38% 5.6Is 1.92 Treebank 9342% 2.26s 1.76 PSD 10.85% 9.97s 2.18 Shakespeare 1.96% 0.07s 2.10 neaten 0.0029 0.42s 1.7828
28 Effectiveness of the SIT Data Source Node Reduction Load Time Node Selection Acceleration XMark 1.64% 0.67s 2.15 OMIM 0.24% 0.07s 2.16 DBLP 0.04% 1.62s 2.11 SwissProt 28.38% 5.61s 1.92 Treebank 93.42% 2.26s 1.76 PSD 10.85% 9.97s 2.18 Shakespeare 1.96% 0.07s 2.10 Lineitem 0.002% 0.42s 1.78
Effectiveness of the sit Index Size: less than 1% of original size Load Time: a fraction of a second Node selection acceleration twice faster than f&B-Index Construction Time: more than 3 times faster than f&B-Index
29 Effectiveness of the SIT ▪ Index Size: less than 1% of original size ▪ Load Time: a fraction of a second ▪ Node Selection Acceleration: twice faster than F&B-Index ▪ Construction Time: more than 3 times faster than F&B-Index
Compression Ratio 口 XQzip+口 XQzip o XMⅲ囗gzip■ XGrind 90 c60 950 40 20 10 XMark DBLP SwissProt Shakespeare Lineitem Data sources XQzip is comparable to XMill and gzip 17 better than xgrind with index size included 42 better than Xgrind without index
30 Compression Ratio 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 100 XM ark OM IM DBLP SwissProt Treebank PSD Shakespeare Lineitem Data Sources Compression Ratio (%) XQzip+ XQzip XMill gzip XGrind XQzip is comparable to XMill and gzip, 17% better than XGrind with index size included, 42% better than XGrind without index