Loading...
Please wait, while we are loading the content...
Projecting network loading of correlated traffic streams under high growth
| Content Provider | Semantic Scholar |
|---|---|
| Author | Wetzel, Timothy D. Lortz, Timothy Thompson, Ashleigh |
| Copyright Year | 2013 |
| Abstract | The optimal design of information technology systems is a multi-objective problem requiring accurate prediction of both normal and peak loading conditions. The calculation of future peak loading requires the projection of average loading as well as a detailed understanding of parameter interactions. The failure to properly account for cross-source correlations for many component systems can lead to a significant underestimation of the total system variability. We present an IT network model that projects both average and peak loading for a system that is expanding traffic at multiple data sources. The average loading is modeled by independently calculating the traffic growth for all data sources. The future cumulative load distribution is calculated by taking into account historical sources of parameter correlations and scaling to account for future average growth. Finally, the duration of the peak states is calculated by simulating a vector Fractional ARIMA time series model. |
| Starting Page | 3967 |
| Ending Page | 3968 |
| Page Count | 2 |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://informs-sim.org/wsc13papers/includes/files/375.pdf |
| Journal | WSC '13 |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
