Guide to VoIP Monitoring, Analysis, and Management
Points to consider when buying a VoIP analysis tool
Protocol support, data capture mechanisms, top list of considerations
When buying VoIP analysis and management tools, here are some of the key product differentiators to keep in mind:
VoIP protocols supported: No one product will track them all, but tools are gaining an increasing range. Protocols supported include Session Initiation Protocol (SIP), Skinny Call Control Protocol (SCCP), H.323, Megaco and Media Gateway Control Protocol (MGCP).
Real-time monitoring or data capture: There are two main ways of providing VoIP analysis - through real-time monitoring, which views data on the fly, or through data-capture, which analyzes data after the fact. Real-time monitoring shows current and/or summarized statistics on active VoIP systems and best serves as a sort of security guard, watching over VoIP traffic as it moves across the wire.
Data-capture and analysis methods record network statistics on VoIP traffic and save them for detailed review through tools designed to enhance the analysis process. Filters – the selection criteria for catching and storing VoIP data for subsequent analysis – provide focus for the examination of the statistics. Many products offer a mixture of real-time monitoring and data-capture processes.
Passive or active: VoIP tools can operate in one of two modes. They can be passive and simply listen to network traffic, gathering statistics in real-time (or near real-time), or actively generate traffic and then gather and compare the results of the test transmissions. The passive approach doesn't add network overhead and can monitor the network for prolonged periods. While active tools can gauge results precisely by comparing packets created and transmitted with the results captured at the far end of the transmission.
Agent or agentless? Some tools require software agents to be installed at network-monitoring points to report VoIP traffic between the monitored sites back to a central administrative console.
GUI considerations: Graphic user interfaces vary widely. Look for:
• An intuitive, linear interface that provides efficient access to information.
• The ability to view a wide range of network activity (the more the better).
• The capability to drill down for greater detail without opening multiple separate windows in a desktop-type interface and, better yet, fix a reported problem.
• The ability to display the actual sound waveform for both sides of a call, allowing visual analysis of problems between the various devices.
• Command-line interface for script execution.
Ease of maintenance: Is it possible to monitor the entire network from a single site? Sometimes additional software probes might be necessary to overcome restricted boundaries. They are installed on the remote hosts. Distributed tools may allow for multiple sites to be simultaneously monitored, either individually or in aggregate.
Types of problems detected: Examine the product's ability to automatically detect problems such as loss of a gateway, controller, WAN link or specific endpoint, and detect call-quality degradation in latency, packet loss and mean opinion score call-quality level.
Sophistication of help and analysis support: Do reports and quality-assurance threshold alerts contain links to background information to assist in explaining the contents?
Customizable triggers/thresholds: A trigger is a set of predetermined conditions that will start the capture of session data automatically. Triggers can be set for pre- and post-event actions on a given threshold or level of session activity. The finer degree of control the better.
Reporting: Does the tool provide preformatted reports of system activity? Can they be customized? How about provide links to third-party reporting tools, such as Crystal Reports?