Cross Platform Data Conversion

 Product Links

File Conversion Considerations
for transmitting files between
IBM Mainframe and UNIX systems

	Application systems are evolving, often moving away from traditional IBM mainframe platforms. In some cases entire isolated applications are being moved to smaller platforms running UNIX.
	More often the application being moved is not completely isolated, but periodically requires data from other mainframe applications. Sometimes data must be routinely sent back to the mainframe as well.

Moving an Application to a Smaller Platform

Industry experience shows that typical COBOL applications are easily ported to smaller platforms. Compilers are available for most platforms. You can easily move program source using the standard FTP utility, which automatically handles the EBCDIC-ASCII translation required for character data. In some cases you can replace user-written applications by pre-packaged software or reengineer them using modern database systems.

Once you have ported an application, you must also port the data files used by that application before production processing can begin. If the ported application must share data files with other applications running on the mainframe, then you must create ongoing processes to port the required files back and forth. A generalized file conversion tool such as netCONVERT is the best way to address these needs.

Data Files are More Complex than Program Source Files

You can move simple data files consisting of fixed length character data using the FTP utility that was used to transfer the program source code. But most applications have much more complex record formats than that.

• Most record layouts include multiple field types, with character and numeric data intermixed

• Different encoding techniques are used for various field types

  Field Type	Mainframe Encoding	UNIX or NT Encoding
  character	EBCDIC encoding	ASCII encoding
  numeric packed	IBM standard encoding	simulated IBM encoding
  numeric binary	big endian	big or little endian
  numeric zoned	IBM standard encoding	no equivalent
  numeric floating point	IBM format	IEEE format

  Records containing intermixed fields with different encoding techniques present a difficult conversion obstacle. And yet a file conversion tool that addresses just the above issues meets only the minimum requirements of most organizations.

Complex Data Files Require Advanced Conversion Facilities

Experience shows that there are additional file format considerations requiring advanced capabilities:

• Complex variable length record formats can be difficult to transfer. RECFM=V, VB, and VBS files present unique challenges to existing FTP facilities. The only general solution is to run the file conversion software on the mainframe so that the variable length records can be accessed directly. You can then transfer the converted file using binary FTP with no further translation.

• Some files contain more than one type of record. Such files also present a difficult challenge. Less advanced file translation solutions cannot discriminate between the various record layouts present. Less complete solutions cannot convert such a file without its first being separated into several smaller files that each contain a single record type. Each separate file requires conversion, followed by a merge and possibly a sort of the translated records. You can avoid all these steps with a more complete file conversion solution such as netCONVERT.

• More complex data stores may require customized support. For example, many organizations use various types of VSAM files to store large volumes of application data in an efficient manner. The ability to directly read and write advanced data stores such as VSAM simplifies the conversion of such data.

A Complete Solution must be a Multi-Platform Solution

Applications running on various platforms on the network must access the data they need without assistance from programs running on other platforms. FTP facilities and Network File Systems are available on all platforms to support this requirement. For a complete solution, you must also have the conversion software available on all platforms.

Less complete file translation facilities are only available on a limited number of platforms. This limitation leads to convoluted time-dependent batch processes where one system deposits a file for eventual use by another.

You can run netCONVERT on all the platforms that are important to your organization, including:

• IBM MVS and VM mainframe systems
• UNIX variants such as AIX, HP-UX, Solaris

netCONVERT - The Complete File Conversion Solution

netCONVERT's combination of advanced file translation capabilities, multiple platform availability, and JAVA-based GUI adds up to a superior data conversion solution. You can port application data files with ease, readily converting them for use by ported COBOL applications. Or translate the data to a "comma separated values" format acceptable to most pre-packaged applications and advanced database systems.

Your data can travel in either direction, with the conversion process running on the most convenient platform for your application batch stream. And netCONVERT is capable of handling your most complex data files. netCONVERT provides a crucial piece of your platform migration strategy.