Reading a computer tape has always been an interesting challenge. It is easy enough if the tape you want to read was created on the machine you are using, but what if that tape came from another machine maybe even another platform altogether? Well if the format was an “interchange standard” then you were in luck..otherwise you had a problem.
The traditional approach for most data centers has been three-fold; custom write code to read the handful of format variations they needed to process – usually an expensive option, with results limited incapabilities, outsource it to a specialist (like ourselves), or refuse to accept data unless it was in one of a small handful of predefined formats.
A number of years ago we created the first real solution for this problem, Media Merge/PC. MM/PC has a common interface and month by month we have added more and more formats to the systems core. The system was designed to mount a tape and automatically determine how the tape was created and decode it. This is no small feat and took a tremendous amount of effort and vision from our development team.
To understand why this is difficult, one needs to understand a little more about what a “format” is.
A format is a layout of data on tape, with a predefined structure that is sometimes static (fixed length records) and sometimes dynamic (variable blocked records, containing variable record lengthed data). A record may be something as simple as an 80 byte string that contains your name, social security #, date of birth, and your last hospital treatment code & its billing cost. A record can also be literally millions of bytes long occupying hundreds of feet of tape.
In all cases the computer has to know what is on the tape and what to do with the bits and bytes it is extracting from the media. Software has to be written instructing the computer how to interpret the data and what to do with it once it has been decoded. Having software which can look at data layouts and automatically determine what it is looking at, is therefore understandably very tricky to write. At eMag we have code that can interpret almost 1800 different tape formats. A large number of these are Mainframe generated.
Mainframes have used a combination of tapes and hard disks as storage devices for over 40 years. Until recently it was not cost effective to store large amounts of data on physical hard disk. For some users who generate Terabytes of data daily it STILL is not viable to store data on disk and so tape has been the preferred storage medium because of its low price and high capacity. Other users have migrated to Virtual Tape with solutions like VTS that use a combination of disk (as cache) and tape.
It is fair to say that Mainframes have generated the vast majority of the tapes in use today. However, today computer usage patterns have changed. The UNIX operating system has slowly but surely grabbed a significant market share within most non-office computing environments. UNIX systems are inexpensive and are very powerful. The net effect has been that applications that were once Mainframe based have now migrated to less expensive but just as capable UNIX systems. The traditional role of the Mainframe has drastically changed, with a number of centers hanging on to their machines just so that they can run their legacy applications.
Tape I/O is a classic example of a legacy operation. If you wanted to read an IBM OS/370 tape on a UNIX platform, it could not be done. The software needed to decode these old legacy formats, until now, has not generally existed on any UNIX system. As of today, this has changed.
We have taken the core of our very successful MediaMerge/PC code, and written a new UNIX GUI for it and we can now handle all the classic legacy formats and read them right into your UNIX system. Our new product, MediaMerge/UNIX is an industry first. There is no other product like this currently available on the market.
Reaction to the availability of this recently announced package has been both amazing and very interesting. The first reaction from the larger data centers has been “Can we use this software to help us get rid of our old Mainframes” – the answer being YES! Other customers are taking advantage of the fact that they can immediately process tape based data much faster using this sytem, due to the inherent processing power of the hardware, than they can using traditional PC’s. Consequently we are seeing customers both step down from Mainframes and step up from PC’s to do their tape processing.
The package itself can run both interactively and in a type of batch sense with scripts rather than Operators providing the usage commands. The menu driven GUI with its context sensitive help and internal Wizard makes the package very easy to use. The software is also loaded with features that allow you to view, test, initialize, copy, and scan tapes. We also have a special extension, RR/UNIX, that allows record-based bit & byte level data manipulation which is a very powerful feature.
Recent Comments