Table 2.
Application | Environment | Repository | Triples | Storage | Query Time |
GORouter | Dual processors of 1.66 GHz, 2 GB RAM | Oracle 10g NDM | 5.5 M | Disk + Memory | 0.74 s |
AllegroGraph | Dual processors of 1.8 GHz, 16 GB RAM | AllegroGraph | 6.88 M | Disk | 172 s |
YeastHub | Dual processors of 2 GHz, 2 GB RAM | Sesame | 1.4 M | Memory | 38 s |
Performance analysis of three semantic web prototype applications:GORouter, AllegroGraph and YeastHub. In this table, the AllegroGraph http://www.franz.com/products/allegrograph/ project used the disk approach to query "LUBM50, Lehigh U. Benchmark" datasets (about 6.88 million triples). The retrieval speed is about 0.04 million triples per second. The YeastHub http://yeasthub.gersteinlab.org project uses the main memory approach to query the UniProt data file (about 1.4 million triples).