Elastic COBOL can operate "somewhat" within the EBCDIC environment. Java runs in Unicode, not ASCII or EBCDIC, character strings within a Java program are Unicode. Elastic COBOL performs data type conversion such that file I/O can be converted from EBCDIC into Unicode. This uses Java's codepage mechanism. The Elastic COBOL compiler option -dt:ibm will compile programs to use this "native EBCDIC" file format and the runtime option ibm.codepage=Cp1047 is used to specify a specific EBCDIC code page when -dt:ibm is in effect.
But Java strings are always Unicode and comparisons are done that way. When a COBOL element (e.g,, PIC X(10)) is assigned a value (e.g., "MOVE 'abc' to MYPICX") the variable contains essentially Unicode single-byte characters, which is ASCII. A check for X'40' in the program would not test for "space". Similarly, internal sorting of characters will be Unicode/ASCII order instead of EBCDIC.
There is also a capability to read/write physical sequential files in IBM variable record formats (RECFM=V or RECFM=VB) with "block descriptor words" and "record descriptor words". This is done with the Elasstic COBOL File Protocols that may be prepended to a file name. For example, "ASSIGN TO 'ibmvb:/home/mark/myfile.dat' " in a COBOL file descriptor will use BDWs and RDWs. These capabilities are available in Elastic Batch Platform (JES/JCL) with the PROTOCOL=IBMVB keyword on the DD statement.
See ecobol Options for more info.