Replication Methods

Enterprise Server automatically uses the replication method most appropriate for your hardware configuration and the data set type. You can also choose a replication method in the Server MCP Disk Properties Dialog Box, provided your hardware and data support it.

This hardware and data set type

Uses this replication method

ClearPath PC or LX computer using a Logical Disk (implemented as a normal Windows file)

-or-

Windows Server and MCP disks on a Storage Area Network (SAN), where the MCP disks are visible to both the MCP and Windows environment (for example, EMC or SCI disks).

Direct disk (with or without caching)

This method gives Enterprise Server direct access to the DMSII data, bypassing the network traffic between MCP, the DBServer, and Enterprise Server. After requesting basic file attribute information from the Databridge Engine, Enterprise Server reads data set disk files directly from the MCP disks without using any mainframe resources.

NOTE:If MCP disks don't appear in the Enterprise Server window under the Enterprise Server system, see Detecting Visible MCP Disks.

 

 

Windows Server (no access to MCP disks or Logical disks)

Disk-region (Indirect disk)

This method allows Enterprise Server to read disk regions like Direct Disk. However, because DBServer is used to access the disks and DMSII data must be retrieved across the network, it's significantly more resource intensive.

 

 

Any of the previous hardware configurations when replicating the following types of data sets without a user-written DLL:

  • Virtual data sets

  • Embedded standard data sets

  • Embedded compact data sets

  • Altered data sets that require a reformat

    NOTE:

    • FileXtract data sets are treated as virtual data sets because they require a Reader program to supply the records.

    • Virtual and altered data sets can be replicated directly in Enterprise Server if a user-written DLL is available for them. See User Written Library Overview.

Standard

Requires that DBServer provide the data record-by-record because some reformatting or other procedures must be done on the mainframe before the record can be sent.