MAINFRAME INFORMATION REPRESENTATION AND STORAGE - IBM Mainframe

IBM Mainframes are unique in some of the terms they use to describe the electronic storage of information. The major components of any digital computer systems are its memory and the CPU. The memory stores information and the CPU manipulates this information and makes decisions based on it.

Programs are logical sequences of instructions that reside in the memory and tell the CPU what actions to take. When the electronically coded information is not required to be in the memory it is stored outside the memory. The IBM Mainframe supports hundreds of information storage devices such as disks, tape drives, etc. These devices connect to memory through small auxiliary computers called channel processors. Each channel deals with a group of information storage devices and relieves some of the I/O load from the CPU.

Billions of tiny transistors form the memory of the computer. Each transistor will or will not be carrying an electric charge. Each transistor stores one bit, binary or digit. Since one bit can represent only two different things, in order to represent the numerous kinds of information like the alphabets, the digits, the symbols, etc., we group the bits together to represent the different characters. Characters formed with 8 bits are called bytes. Because we use 8 bits to store each character of data to store each character of data a total of 256 possible bit patterns exists. In IBM Mainframes the bit patterns are mapped to the language symbols using a scheme called EBCDIC, which is different from the ASCII scheme.

If you group characters together to store a customer name, address and other items of information, each group of characters is called a field. If you group together fields to collect the information about an entity say a customer or an employee you form a record. The MVS can handle records up to 32, 760 bytes. You can place records in the computer memory using a table. If you copy such records out of memory and house them on a disk or tape you get a file.

In the IBM Mainframe terminology, you call a file a dataset. A dataset is a collection of record. In a Mainframe system there will be thousand of datasets. Each dataset should have a unique name. MVS support very large DataSet Names (DSN). A DSN can be as long as 44 characters in the following format: xxxxxxxx. xxxxxxxx. xxxxxxxx. xxxxxxxx. xxxxxxxx

You can compose DSNs using groups up to 8 characters. Each group should start with a capital letter and the next seven characters can be either letters or numbers. You can also use the symbols @, $, and # in DSNs. MVS keeps track of groups of datasets by referring to the first part of the names, called the qualifiers. When you use TSO/ISPF, the first part of the DSNs will be usually your user ID or group ID.

For example, I was working for a project called SEGA and my user ID was SEGAWL and I used to store all my COBOL source code in the dataset named SECAWLSEC.COBOL, my personal files in SEGAWL.AWL.PERSNL, etc. It is not necessary to use all the 44 characters or all the 8 characters in the group.

Mainframe security software controls the access to datasets by referring to the first part of the DSN. Using either TSO or JCL you will be able to access only datasets starting with certain names as permitted by your installation's security mechanism.


All rights reserved © 2018 Wisdom IT Services India Pvt. Ltd DMCA.com Protection Status

IBM Mainframe Topics