SYNCSORT: SUM FIELDS

SYNCSORT is one of the more powerful tools on the mainframe. One of its features is it can sum up a specific column in your file based on some keys. Say you have a file that contains items sold and number of items sold per transaction as seen below:

z/OS Compiling Programs that Use DB2 or CICS

In our previous article: z/OS Compile Process, we talked of the basic Compile process. On the mainframe though, you can write programs that call other subsystems, like a database system (ie. DB2) or a transaction system (ie. CICS).

Auditing RACF Using Microsoft Access or any SQL Database

Executive Summary

Aside from being difficult, auditing RACF requires special skills. It needs understanding of RACF concepts plus understanding of the other subsystems being audited. A comprehensive manual audit took four auditors two months to complete. This can make the audit report irrelevant by the time it is released because errors may have been corrected already - making the auditors look incompetent - or worse, security breach may have occurred without the auditors detecting it.

SYNCSORT Advanced Features

Many use SYNCSORT for sorting files. But SYNCSORT can do a lot more. It can be used to format records. Because of this feature, it can be used for other purposes. (These features are also available in DFSORT, although some features work differently in DFSORT).

z/OS File System

Every operating system has its own file system. z/OS is no exception. Windows and *NIX file systems are based on hierarchical directory structure. Basically, you have one root directory and you can create directories (sometimes called folders) or files under that root directory. Mainframe disk files are very different to this hierarchical approach.

Unix System Services (USS)

Probably unknown to many in the industry, the IBM mainframe runs Unix System Services (USS) with the z/OS operating system. USS runs as a separate address space in z/OS as a started task.

Types of Mainframe Processing: Batch and Online

There are two types of mainframe processes:

1. Batch Processing
2. Online Transaction Processing

Mainframe Programmers: Where are they?

One of the difficulties in working on a mainframe environment is looking for good people. A lot of those who know mainframes are retiring or have retired and are living on their boat. For a lot of those in university, mention the word mainframe and they will run away as if the mainframe is a plague that should be avoided.

Mainframes: Are they Boring?

I started my career as an operations analyst for MVS. I later moved on to be a systems programmer for MVS and IMS. This was and still is the best part of my career.

I had since graduated from MVS to pursue other areas of IT. The past few years brought me back to the mainframe. I was really surprised to see a lot of improvements on MVS. You can use it to deliver web services. You can connect it to any web application, access DB2 and IMS from a web application.

I left MVS when MVS/XP was coming out. I was not able to play with 31 bit addressing.

In the past several years, things have changed. Real storage is now in Gigabytes, whereas before, 16 MB was the highest an application can go - and that was in virtual storage!

You now have FTP, SFTP and FTPS as transport protocols.

There's also OMVS which allows you to use *NIX services on MVS!

The application architecture has changed also. There are now technologies that allow you to wrap existing IMS and CICS transactions into web services that can be served to web clients. You can call API's that allow your MVS COBOL code to access web services outside of MVS. There are also technologies that allow you to access your hierarchical IMS databases via SQL. And lastly, you can present your dull green screens as a web page.

One thing that I realize is these new buzzwords: like virtualization, XML, even the concept of paging, all these have their origins on the mainframe.

IBM allowed virtual machines through their VM product. With XML, data is presented hierarchically - and IMS started this concept. IMS is a lot more powerful though. The concept of separating data from presentation also has its origin from the mainframe. With MFS, IMS programs did not need to concern themselves with how the data is presented. The presentation is coded in MFS.

Lastly, paging - OS/VS1 had this. That was what VS stood for - Virtual Storage meaning it provided more memory than what it really had. The advantage of how the mainframe implemented VS is the way VS is handled. For Windows and *Nix systems, VS is one big block of memory shared by all the tasks. on the Mainframe, each address space has the same amount of virtual storage. So for 31 bit addressing, each address space has 1GB of virtual memory. It does share some overhead with the system but from its perspective, it has the whole 1GB to itself. Of course, when you have 64bit addressing, the available virtual storage is a lot more.

Now who says mainframe technology is dying and boring?