Data model for trader voice recordings

The system stores recordings in the database using 2 different data models:

Standard data model: a single conversation entry (CDR) is linked to a single media file (1-to-1 relationship), suitable for normal phone calls, instant messages, files, etc.
Trader voice data model: conversation entries can be linked to multiple media files (Many-to-Many relationship), suitable for trader voice recordings

The Many-to-Many relation, used by the trader voice specific data model, is supported by 2 types of records in the database:

CDR-only: CDR-only records contain CDR information and reference to one or more media-only records
Media-only: One media-only record is shared across multiple CDR-only records (channel mixing, TPO recording)

.

The trader voice data model is available for the following trader voice integrations:

IPC Unigy
BT IPTrade
Speakerbus

This data model allows storing a single copy of the media in case of mixed recording channels or TPO based recording where multiple calls/sessions are referencing a single media entry.

The trader voice specific data model is enabled by the default for the supported integrations but can be optionally disabled. If the trader voice specific data model is disabled, the system will use the standard data model for trader voice recordings. In case of open lines in a mixed recording channel, it means that each CDR will have a separate media file entry containing the same data.

Generally, there is no difference between the data models when it comes to features, although certain features work slightly differently when the system uses the trader voice specific data model. The following table describes the effect of the trader voice specific data model on key features:

Feature	Operationfor trader voice data model
Search	CDR-only (and standard) records are shown by default, the user can switch to show media-only records as well
Playback	Requires post processing to stitch and cut media-only records to CDR-only records. Seamlessly integrated intoweb basedplayback, media files are generated temporarily and cached
Data retention	Set for both CDR-only and media-only records in Verba On WORM storages: only media-only records have retention period setting
Deletion	Media-only records have their own retention but even so, cannot be deleted until a CDR-only record references it
Voice quality check and transcoding	Applies to media-only records
Transcription, phonetic indexing	Supported for CDR-only records using media stitching
Export	Calls can be exported either with stitched media (standard model) or as per turret model
Import	Both standard and turret model is supported
Desktop/agent screen recording	Not supported