In a recent application xGT achieved near linear scalability when querying 3 terabytes of cyber data with 20 billion graph edges and 212 billion edge properties. On another benchmark, the tool demonstrated orders of magnitude improvements in speed, reducing query time from 179 hours to 12 minutes.
Extreme parallelism allows the tool to schedule work efficiently across thousands of cores. Trovares’ engineers have spent years perfecting multithreaded code for supercomputing. For xGT they brought parallelized code from high performance computing to graph analytics.
While traversing trillions of edges, xGT uses optimization to pare the search space for greatest efficiency. By quickly removing irrelevant data from searches, xGT queries are multiple times faster than non-optimized searches. Parallelism and optimization produce search results hundreds of times faster than conventional graph analytics tools.
Parallelism and Optimization for Speed and Scale
Cyber and Fraud
xGT is in use today by the Department of Defense. It’s ideally suited for applications with large datasets and time constraints such as cybersecurity, fraud detection, and graph analytics.
xGT surfaces characteristics of large networks in seconds. The visualization of xGT query results was produced using the Graphistry data visualization tool which speeds pattern discovery.