How does Adaptive Monitoring Work

Owned by Ramya Raman
Last updated: Aug 13, 2024 by Danna Mann
3 min read

The TAG MCM-9000 has two modes of operation:


• Full monitoring mode where the input source is fully decoded and can be displayed in real-time in a Tile of the Mosaic Output while probed, analyzed, and alarmed for errors.

• Partial monitoring mode where the input source is probed, analyzed, and alarmed for errors, but the video is not fully decoded and CAN NOT be displayed in real-time in a Tile of the Mosaic Output.

The second option, which we also call “Monitor Only Mode” is achieved by not decoding every frame of the video, but only part of the frames (in short intervals between each decoded frame), enough to detect the presence of the video, but not to display it in real-time.

The benefit of this mode is that it dramatically reduces the load of the Server CPU, using about 1/3 of the resources needed to fully decode, monitor, and display an input source in a Tile.

Until now, a Server could operate in either Full Monitoring Mode, or Monitor Only Mode, so depending on the quantity and type of sources, a configuration setup had to be done with Servers operating in either one of the modes ahead and for the duration of the deployment.

TAG’s New Adaptive Monitoring

The new “Adaptive Monitoring” feature allows operating in a Hybrid mode where the selection of which mode to operate in is made per Input Source, and each Server can be dynamically and adaptively assigned to monitor any combination of the two operating modes.

The constraint to set which mode each Server is operating in is now alleviated, especially in the Cloud deployment where Multicast support is not available for the most part, so combined deployment of Full Monitoring and Monitor Only modes is not supported in the same deployment.

Adaptive Monitoring Capabilities

The essence of the new Adaptive Monitoring feature is that it enables the user to choose per channel the monitoring mode: Full, Light, or Extra Light, with the ability to change it dynamically.

In the Light and Extra Light modes, the system does full-time monitoring and alarming on all aspects of the source, including the Audio, time intervals, Video decoding, and video monitoring, as in the current “Monitor Only Mode” (alarming continues to be full time).

The Extra Light mode has longer intervals before decoding a full Video frame than the Light mode.

As for CPU resources for the different monitoring modes:

  • Full Monitoring : Requires one full license and regular Hardware usage (SD Points depending on the source)

  • Light Monitoring : Requires one full license and 0.33 (1/3) HW usage (SD Points)

  • Extra Light Monitoring : Requires one full license and 0.2 (1/5) HW usage (SD Points)

Light / Extra Light Monitoring Functionality

When a channel is in Light or Extra Light mode and is assigned to the Mosaic Tile by either a Manual, API command, or CED-triggered need, and there are Hardware resources available in the Server, the channel is monitored or needs to be displayed on, the monitoring is upgraded to Full mode until further change, and it can be displayed on a Mosaic Tile. If no hardware resources are available on the Monitoring Server, an error icon is displayed on the Mosaic Tile, indicating no hardware resources to perform the desired action.

The active monitoring mode of each channel is shown in the GUI statistics and available by the API.

This mode is not available or relevant for uncompressed channels since there is no decoding of compressed content. CPU resources are always the same, and the Live Production application needs to see real-time live video on the Mosaic Tiles.
It is, however, available for all compressed formats MPEG-TS, Zixi, J2K, OTT, etc.

When source is using Light mode, the transport layer and audio are fully monitored, as if it was full mode.

Video is monitored as following:

  1. Decoding starts with the first I frame and go for about 20-30 frames.

  2. Decoding stops for 3 seconds.

  3. After 3 seconds decoding is restarted, awaiting for the next I frame to begin (Interval might be slightly longer than 3 seconds).

  4. Once the I frame is received, decoding starts and go for another 20-30 frames.

  5. Steps 1-4 is repeated.

Extra-Light mode works the same way except for the decoding stop interval which is 5 seconds (step 2).