Medianet Reference Guide - Cisco

Network-Embedded Management Functionality
Chapter 6
Medianet Management and Visibility Design Considerations
Transport Protocol (RTP) traffic as “Other UDP “or “VoIP”, because RTP can use a range of UDP ports.
Further, the ability to drill down to monitor and generate reports that show the specific hosts and flows
that constitute RTP video and/or VoIP traffic, versus all UDP flows, may be limited. Further, both VoD,
and in some cases video surveillance traffic, can be sent using HTTP instead of RTP. Therefore,
generating useful reports showing medianet-relevant information, such as how much video data (RTPand/or
HTTP-based) is crossing a particular point within the network, may not be straightforward.
For devices such as TelePresence endpoints and IP video surveillance cameras, you can often simply
assume that most of the data generated from the device is video traffic, and therefore use the overall
amount of IP traffic from the device as a good estimate of the overall amount of video traffic generated
by the device. Figure 6-2 shows a sample screen capture from a generic NetFlow collector, showing flow
information from Cisco TelePresence System endpoints to a Cisco TelePresence Multipoint Switch, in a
multipoint call.
Figure 6-2
Sample Host Level Reporting From a NetFlow Collector Showing TelePresence
Endpoints
Note
Figure 6-2 shows a screen capture from the open source ntop NetFlow collector.
The IP addresses of the TelePresence devices have been replaced by a hostname to more easily identify
the endpoints. As can be seen, both the actual traffic sent and received, in terms of bytes, as well as the
percentage of the overall traffic seen across this particular interface over time are recorded. Such
information may be useful from the perspective of determining whether the percentage of bandwidth
allocated for TelePresence calls relative to other traffic, across the interfaces of this particular collection
point, matches the actual data flows captured over an extended period of time. However, this information
must also be used with caution. Flow records are exported by NetFlow based on the following:
• The flow transport has completed; for example, when a FIN or RST is seen in a TCP connection.
• The flow cache has become full. The cache default size is typically 64 K flow cache entries on
Cisco IOS platforms. This can typically be changed to between 1024 and 524,288 entries.
• A flow becomes inactive. By default on Cisco IOS platforms, a flow unaltered in the last 15 seconds
is classified as inactive. This can typically be set between 10 and 600 seconds.
• An active flow has been monitored for a specified number of minutes. By default on Cisco IOS
platforms, active flows are flushed from the cache when they have been monitored for 30 minutes.
You can configure the interval for the active timer between 1 and 60 minutes.
6-8
Medianet Reference Guide
OL-22201-01