1,083 time series recording hourly maximum traffic flow across 128 systems and 16 logic data centers. The data covers a period exceeding 4 months and is partitioned into training and testing segments with a 32:1 ratio.
Use Cases
- Train forecasting models to predict hourly maximum traffic flow based on historical system patterns
- Evaluate anomaly detection algorithms on 1,083 time series to identify infrastructure irregularities
- Analyze traffic correlations between 128 systems and 16 logic data centers to optimize resource allocation
Strengths
- 1,083 distinct time series representing system-level traffic
- Hourly maximum traffic flow metrics collected over a 4-month period
- Infrastructure coverage across 128 systems and 16 logic data centers
- Standardized training and testing split ratio of 32:1