NxTran

Mainframe modernization presents significant challenges for organizations seeking to reduce costs and improve system efficiency while maintaining critical applications. This technical article examines NxTran, a comprehensive mainframe modernization service developed by Kumaran Systems, which enables automated migration to distributed platforms while preserving application functionality. Through detailed analysis of the migration process, data transformation techniques, and integration capabilities, we uncover how NxTran achieves significant reductions in operational costs and time to market for Fortune companies, government agencies, and large corporations.

Tool Overview and Development

NyTran, developed by Kumaran Systems, facilitates mission-critical application migration to distributed platforms while reducing MIPS usage and IT costs. The company's mainframe modernization service automates this process through NxTran, retaining application functionality while achieving significant reductions in operational costs and time to market. Over 2000 successful migrations have been executed for Fortune companies, government agencies, and large corporations, demonstrating the service's reliability and scalability.

The modernization process begins with NxTran's proprietary tool suite, which employs advanced techniques to automate migration while maintaining application integrity. By leveraging process efficiency protocols and accelerators developed through extensive experience across various enterprise sectors, Kumaran ensures reliable service delivery. The company maintains ISO 27001:2013 Information Security Management System certification, demonstrating a strong commitment to client security and quality standards.

Kumaran Systems stands as a partner to its clients rather than a vendor, addressing potential concerns about vendor lock-in through open-source extensions and emulators that maintain data ownership while enabling cross-platform compatibility. The company's expertise extends across multiple industries, including banking, insurance, manufacturing, retail, transportation/logistics, education, healthcare, public sector, and telecom, with proven success in delivering comprehensive application modernization services.

Modernization Process and Benefits

The modernization process leverages NxTran's proprietary tool suite to automate migration while maintaining application functionality, achieving significant reductions in operational costs and time to market. The service employs sophisticated techniques to reduce field complexity and optimize data structure definitions, ensuring efficient translation between mainframe and distributed platforms.

Data migration begins with the generation of COBOL and JCL source files from VSAM files. The COBOL source sequentially reads VSAM files and writes physical sequential files, while same-platform conversions use IDCAMS (REPRO) with VSAM structure repository inputs. Character changes between EBCDIC and target HOST formats are performed using DB2 load utility, which efficiently handles all data types including large objects (LOB) and user-defined types (UDT).

For database integration, NxTran generates Data Definition Language (DDL) statements based on the application schema model. These statements create necessary tables and structures when executed against the target DB2 database instance. The DDL process enables seamless data upload from physical sequential or flat files without triggering application constraints, except for index uniqueness validation.

The modernization approach supports complex integrations with CICS (Common Integration and Control System), a general-purpose data communication platform. CICS manages telecommunication between application programs and remote/local terminals/subsystems while controlling concurrent program execution and data access transactions across multiple systems.

Data Transformation Techniques

NxTran employs sophisticated data transformation techniques to enable cross-platform migration while retaining application functionality. The tool's methods include field splitting, redefinition handling, and multi-dimensional array conversion, with specific processes for managing complex data structures.

Field splitting divides multi-dimensional fields into multiple files. For example, a personnel file with id, forename, and surname fields is split into two record types: personnel (id integer, forename char(10), surname char(20)) and personnel_mx (id integer, indx integer, address char(20)). Program access remains consistent, with personnel.address[1] accessed through personnel_mx records where id = personnel.id and indx = 1.

Data redefinition is handled through a process of reducing fields to their lowest common denominators and replacing subsequent entries. For example, a personnel data structure with a name field (PIC X(30)) and detailed_names field (REDEFINES name) containing 07 forename and 07 surname subfields would be simplified to separate id (5), forename (10), and surname (20) fields. NxTran provides two options for managing incompatible redefinitions: ignoring the specific redefinition for memory reduction purposes or requiring manual intervention to clarify the record structure.

The tool's array handling capability splits multi-dimensional arrays into two files. For instance, a personnel file with address fields occurring three times would be converted into two record types: personnel (id Integer, forename char (10), surname char (20)) and personnel_mx (id Integer, indx integer, address char (20)). This approach ensures that program access remains consistent across the transformed data structure.

NxTran's data conversion processes generate COBOL and JCL source files using a VSAM file repository. For cross-platform conversion, the tool reads VSAM files sequentially and writes to Physical Sequential files, performing character changes from EBCDIC to prepare the data for transfer to the target HOST. Same-platform conversions utilize IDCAMS (REPRO) with inputs from the VSAM structure repository.

The tool's data upload process uses the DB2 load utility to move data from physical sequential or flat files into the target DB2 database. This utility handles all data types, including large objects (LOB) and user-defined types (UDT), while performing no trigger firing or referential constraint checking except for index uniqueness validation. This approach provides exponentially faster load times compared to traditional methods.

Technical Implementation

The technical implementation of the migration process relies on a detailed repository of VSAM file structures, built through parsing COBOL source files and CICS metadata. For COBOL programs (CASE 1), NxTran analyzes file structures from source files or COPYBOOK files, resolving any indirect references through JCL statements that link logical file names to physical device names. The tool then updates the VSAM structure repository with accurate cluster names and definitions.

In the case of CICS/COBOL applications (CASE 2), NxTran leverages CICS's abstraction layer. The tool accesses metadata through the File Control Table (FCT) and Symbolic Map Files, which manage how application programs interact with VSAM data. This approach allows application programmers to focus on business logic rather than file management details.

The VSAM structure repository serves as the foundation for building a normalized database schema. The tool transforms this repository into a Third Normal Form (3NF) data model, where each VSAM file becomes a corresponding DB2 table. The base VSAM key serves as the primary record identifier, while COBOL copybook fields map directly to DB2 columns.

For data type translation, NxTran refers to IBM's official mappings in Appendix A of the mainframe documentation. The tool can extract comprehensive field descriptions from multiple COBOL programs within an application, ensuring accurate type conversions during the schema generation process.

During the data migration phase, the tool generates COBOL and JCL source files that sequentially read VSAM files and write physical sequential files, performing character changes from EBCDIC to Unicode as part of the preparation for DB2 upload. For same-platform conversions, it uses IDCAMS (REPRO) with VSAM structure repository inputs.

Integration Capabilities

CICS (Common Integration and Control System) serves as the primary middleware layer for integrating mainframe applications with distributed systems. This general-purpose data communication platform manages telecommunication between application programs and remote/local terminals/subsystems, while controlling concurrent program execution and data access transactions across multiple systems.

A CICS transaction consists of logically related programs forming multiple tasks, with each transaction creating a new task containing Logical Units of Work (LUWs) - a programming construct similar to atomic transactions in relational database management systems (RDBMS). While tasks are managed by the CICS task control program, multitasking capabilities enable the system to manage multiple tasks simultaneously. This architecture facilitates pseudo-conversational transactions, where a series of non-conversational transactions appears as a single conversational transaction to users.

The J2EE (Java 2 Platform, Enterprise Edition) application model offers a multi-tiered distributed architecture that enables component-based development and deployment across various machine configurations. This component-based approach allows for flexible scalability and maintenance, as application logic can be divided into client-tier components running on client machines or browsers, and web-tier components running on J2EE machines.

The integration between NxTran and these mainframe components enables seamless migration while preserving existing intersystem communication capabilities. NxTran generates COBOL and JCL source files using a VSAM file repository, performing sequential reads of VSAM files and writes to physical sequential files for cross-platform conversion. Character changes between EBCDIC and target HOST formats are handled through a process that prepares data for upload to the target DB2 database using the DB2 load utility.

The tool's data upload process utilizes the DB2 load utility to efficiently move data from physical sequential or flat files into newly created tables or existing ones. This utility supports all data types listed in IBM's official mappings, including large objects (LOB) and user-defined types (UDT), while performing basic referential integrity checks focused on index uniqueness validation. The combination of CICS and J2EE components allows for near real-time processing capabilities while maintaining the existing mainframe's intersystem communication capabilities.