Forecasting has the potential to interfere in decision-making and gives a picture of what can probably be expected. Every business decision depends upon forecast on future trends and hence, it is important to give maximum attention to it.
To plan network resources, timely and accurate prediction of bandwidth is beneficial to avoid network bottlenecks. Also, to handle the ensuing network needs efficiently and to regulate traffic, it is of utmost importance to predict data.
While a time series can be forecast using statistics or machine learning, NetFlow Analyzer employs techniques like autocorrelation, seasonality trend loss decomposition and regression to forecast reports.
A forecast report can only be as accurate and reliable as the data provided. The granularity and accuracy of the forecast differ based on the available data. The below given table speaks on the accuracy of the forecast that would be generated based upon the availability of past data.
Time period | Past data | granularity |
1 week | 36 days | Hourly |
2 weeks | 70 days | Hourly |
1 Month | 130 days | Hourly |
3 Month | 180 days | Daily |
6 Month | 180 days | Daily |
1 Year | 5 years | Weekly |
Click on Reports -> NetFlow -> Forecast to generate the forecast report.
You can select the Source for which you need to generate the forecast.
Traffic Type gives the option to view the traffic graph details based on Speed, Volume or Utilization.
Forecast reports can be generated for a period ranging from 7 days to a year. However, the Forecast period cannot be customized and can only be viewed for a pre-defined set of data.
Business Hour Filters and the Exclude weekends option can also be enabled. By default, Saturday and Sunday are set as weekends.
On enabling the Show History button, the users can view the past trends based on which the forecast has been generated.
You can mail the report or export data as CSV, PDF by clicking on the icons in the top right corner of the view.