Version 26.10

Flussonic 26.10: VOD distribution in Catena, multicast standby in Mcaster and NVR updates with rollback in Watcher

Version 26.10 helps operators keep content available, use their existing subscriber applications and maintain recorders remotely. Here is what changed in Catena, Mcaster, Watcher and Flussonic Media Server.

Here is what we did for Catena

A VOD library can serve viewers from several servers. Set the number of copies in the asset template, and Catena creates them on other servers. New playback requests go to available copies, choosing the server with fewer sessions. This spreads demand for popular content and keeps new requests away from a server that has been marked unavailable. The console shows copy status and viewers per copy. Existing sessions are not transferred automatically: if their server fails, the player must reconnect.

Keep the subscriber applications you already deployed. The compatible IPTV playlist supports applications such as IMPlayer that used Flussonic playlists. Playback links use subscriber credentials rather than short-lived playback tickets, so an imported playlist does not stop working merely because its ticket expired. Channel logos appear in playlists and XMLTV, and groups help viewers navigate the channel list.

Use your existing DRM provider. Catena now obtains encryption keys directly from EZDRM. Operators can keep that service when moving channels from Media Server; the console supports its credentials, and the Flussonic import transfers the supported drm ezdrm settings. This release uses one key for all tracks of a stream.

Choose hardware for each task. A dedicated DVR node records the archive while ingest and transcoding run elsewhere, including on GPU servers without archive storage. This role currently applies to streams without DRM.

Here is what we did for Mcaster

Prepare a backup output that takes over automatically. The standby push listens to its multicast group and waits while another sender is active. After the configured silence interval, it starts sending; the default interval is one second. The console shows whether it is waiting or active. This is output standby, rather than a promise of frame-perfect switching or cluster twincast.

Send each receiver only what it needs. From one quality ladder, send 1080p to one multicast group and 720p to another, selecting the audio tracks for each destination. Track selection is available for MPEG-TS pushes over UDP/RTP, SRT and RIST, and for RTMP. It reduces unnecessary traffic and avoids creating separate input streams for each destination. Track selection for scrambled MPEG-TS streams is not supported in this release.

Fit into the broadcast network. Select the multicast interface directly, especially when management and media use different networks. For ST 2110 installations with compatible hardware and drivers, the optional AF_XDP receive path avoids packet copies through the ordinary network stack.

Return the stream to an SDI broadcast chain. Mcaster now sends raw video and audio through a compatible V4L2 output device. Compressed inputs require an explicitly configured decode step. SDI input can use time-of-day timecode, and the output carries timecode and aspect signalling so an anamorphic SD 16:9 picture is displayed with the right proportions. While an enabled output has no video frames, it keeps the line active with black video and silence.

Preserve ad cues for downstream systems. SCTE-35/104 markers pass through the media path without transcoding, with signalling in MPEG-TS, HLS and DASH. The new ad tab shows announcements, breaks and returns, helping engineers check the cues supplied to external ad insertion systems. The feature transports and monitors cues; it does not insert advertising itself.

Here is what we did for Watcher

Maintain recorders with a clear view of the result. The update page shows downloading, backup, installation and verification stages, their duration and any error. The recorder keeps the last ten configuration snapshots; an administrator can restore a selected snapshot from the interface. If an update fails or is interrupted by a power loss, the updater attempts recovery and can roll back automatically. This reduces the need to visit the site to understand or recover from an update failure.

Find and export the relevant evidence. Detection event markers point to moments that deserve attention in the archive. The download dialog lets the user choose video quality and an audio track, producing a smaller file when full resolution is unnecessary.

Try the new media core on a selected streamer. A pilot migration script checks the installation, prepares the switch from Flussonic to Sapsan, verifies it and supports rollback. The pilot covers cluster media nodes and standalone streamers; all-in-one installations are outside its scope. Separately, cloud playback from Sapsan NVRs now uses the newer inter-server transport for better stability.

Here is what we did for Flussonic Media Server

These fixes address daily operation of existing installations: viewers can watch long recordings, servers retain memory for their workload, and an individual slow RTMP client is handled correctly.