Application assessment

The Application assessment is a process that extracts and collects vital information about the objects being migrated, and provide the informative base for the following steps in the rehosting process.

This is a complex task, that requires either automated tools for the application understanding (i.e. HTWC ICON SUS, or other similar tools) and human interactions to obtain information from the application maintainers.

The information that must be collected are described below.

Application sources

The process must identify the sources that really needs to be migrated. Normally this task is performed with automated tools (like ICON) that, starting from the whole application portfolio, allow to identify and extract the source code currently running on the mainframe, discarding eventually dead branches of code, obsolete sources, etc. thus correctly defining the set of sources to be migrated.

Application boundaries

Once the application sources have been identified, it is also important to identify the application boundaries, that are the points of contact of the identified set with other applications or systems. This bindings may be:

The knowledge of this information is really useful to pre-define a solution for migrated code, thus preventing unexpected problems in the test phase.

Data to migrate

From the analysis of the set of sources you are going to migrate, you must identify the set of files and databases used by the programs. therefore to be migrated.

Remember that not all the files of this set must be migrated, in fact you do not need to migrate those data that can be regenerated on the target system, like temporary files.

Data structures

The data conversion process relies on the knowledge of the structures of data being migrated, that must be identified and provided to the converters.

Structured data, like database tables, do not need further investigation, because the table descriptions contained in the DDL are enough for XCONV that recognizes fully them.

The problem arises for unstructured data, like files and DL/I (IMS/DB) segments. For these items, their record description must be identified (normally a COBOL copybook). This information may be extracted with automated tools, or obtained by the user documentation.

Basically you need to obtain the one-to-one relationship list among files/segments and copybooks.

Furthermore, for multi-record types files and segments, you must isolate the rules that allows to recognize the correct data structure: this information is normally obtained by the application maintainers or by the program coding.