Replicate is happy to read the Active Transaction Log and/or Archived Tlog and extract the changes for selected table from it to apply to a target and/or store in a changes table on some target database. It can find the Archived log location joining backupmediafamily and backupset system table. It can use MS SQL stored procedures to read the file, or read them directly if access is provided. This is a powerful, source transparent, convenient way to find all change on an ongoing, quasi-non-polling, near real time CDC method. It's the INTENDED, and best way to use Replicate. The 'connection' is just a valid logon to the DB using sufficiently privileged username (see pre-requisites).
MS SQL can also serve as Generic ODBC source the the base tables are queried for rows with a 'key' (sequence number, date&time stamp) higher than the last change read. See chapter "8.19 Using ODBC to connect to a source". That's an alternative, non-transparent, discouraged 'connection'.
Replicate is happy to read the Active Transaction Log and/or Archived Tlog and extract the changes for selected table from it to apply to a target and/or store in a changes table on some target database. It can find the Archived log location joining backupmediafamily and backupset system table. It can use MS SQL stored procedures to read the file, or read them directly if access is provided. This is a powerful, source transparent, convenient way to find all change on an ongoing, quasi-non-polling, near real time CDC method. It's the INTENDED, and best way to use Replicate. The 'connection' is just a valid logon to the DB using sufficiently privileged username (see pre-requisites).
MS SQL can also serve as Generic ODBC source the the base tables are queried for rows with a 'key' (sequence number, date&time stamp) higher than the last change read. See chapter "8.19 Using ODBC to connect to a source". That's an alternative, non-transparent, discouraged 'connection'.
