Perform Detailed Monitoring - Real-Time Communication on AWS

Perform Detailed Monitoring

End users of real-time voice and video applications expect the same level of performance as they achieve with traditional telephony services. So, when they experience issues with an application, it ends up hurting the provider’s reputation. To be proactive rather than reactive, it is imperative that detailed monitoring be deployed at every part of the system that serves end users.

Figure 10: Using SIPp to Monitor VoIP Infrastructure

Many open source tools, such as iPerf or SIPp, and VOIPMonitor, are available that can be used to monitor SIP/RTP traffic. In the preceding example, nodes running SIPp in client and server modes are measuring SIP metrics such as Successful Calls and SIP Retransmits between all four US AWS Regions. These metrics can then be exported into Amazon CloudWatch using a custom script. Using CloudWatch, customers can create alarms on these custom metrics based on a certain threshold value. Automatic or manual remediation actions can then be taken based on the state of these CloudWatch alarms.

For customers not wanting to allocate engineering resources needed to develop and maintain a custom monitoring system, many good VoIP monitoring solutions are available on the market, such as ThousandEyes. An example of a remediation action is changing the SIP routing based on increased SIP retransmits.