Analyze Bandwidth Usage Patterns

   Steps:

1. Create a traffic class if one does not already exist for the traffic you want to analyze.
   

   For example, if you want to analyze usage patterns for file transfers, then you probably don't need to do anything because the presence of FTP traffic automatically prompts PacketWise to create FTP classes.
   
   But if you want to analyze file transfers from a certain server, then create a new traffic class under the FTP class that sub-classifies FTP traffic by server (for example, a class called Inbound/FTP/ServerName).
   
   For more information, see Traffic Tree Overview and/or Traffic Classification Overview.
   
   

2. Ensure that PacketWise has had sufficient time to measure usage patterns. If in doubt, wait for at least a couple days. A week is even better.
   
   
3. Select your report tab to display the Network Performance Summary. Change the interval at the top to cover the last week. Find the Top 10 pie charts. Check to see if your application is listed in the highest bandwidth consumers.
   
   Evaluate whether you think it's appropriate that your application be listed or not. Is it a critical application that needs definite bandwidth, such as SAP, Oracle, or a Citrix application? Then it's probably appropriate to be listed. Is it a critical application that uses tiny amounts of bandwidth, such as Telnet? Then it's okay not to be listed. Is it a bandwidth-greedy but superfluous application such as music sharing? Then you'd prefer not to see it listed.
   
   
4. Create a graph of bandwidth utilization with peaks for your traffic class. Adjust the graphing interval to cover one week.
   
   Do the peaks extend near link capacity? Then your class' bandwidth consumption is probably impacting other applications' performance at least occasionally.
   
   Is the average line much lower than the peak line? Then your traffic has sporadic spikes, but the class' typical consumption isn't so impacting.
   
   Does you class' usage correspond to personnel's work schedules? Are their identifiable patterns?
   
   Consult examples of applications' bandwidth utilization graphs and their interpretations.
   
   
5. Compare your application's utilization graph with the one for your whole link on your Network Performance Summary. Do your application's peaks impact the link's peaks?
   
   
6. Examine the Monitor Traffic window. Find the row for your traffic class.
   
   Examine the current (at the point you displayed the monitor screen) rate and the peak rate for your class. In this example, the Inbound class folder on the Monitor Traffic screen shows 3.1M as a Peak value, which is probably two T1 lines. The class for inbound mail never uses more than 1.5 Mbps, so it probably can access only one T1 line.
   
   
   
   
   
   
7. If you want to explore the top contributors and consumers of the traffic in your class, set Top Talkers and/or Top Listeners on your class. Then wait for more traffic to pass and examine the results.
   
   Examples of how Top Talkers (TT) and Top Listeners (TL) can be handy:
   • Use TT on an Oracle traffic class to reveal the most active Oracle servers.
   • Use TL on a traffic class for the three most popular sports sites to reveal your top sports fans.
   • Use TL on the current music-download application's traffic class to reveal those who retrieve the most music.
   • Use TT on the same class to reveal servers most tapped as music sources.
     
     
8. You can compare bandwidth utilization in several classes by graphing multiple classes. Choose the metrics for utilization average and peak.
   
   
   
   
9. If you want to monitor bandwidth usage automatically without needing to check usage manually, the adaptive response feature can be helpful. See Analyze Efficiency with Traffic Performance Agents and Create Partition Utilization Agents.