Mainframes: Time to rip the heart out of your business?

Mainframes: Time to rip the heart out of your business?

Summary: The mainframes at the heart of some companies are decades old, but removing them is a massive ordeal which one HP exec likens to a heart transplant — a necessary yet painful operation many companies are loath to undergo.


The mainframes at the heart of some companies are decades old, but removing them is a massive ordeal which one HP exec likens to a heart transplant — a necessary yet painful operation many companies are loath to undergo.

Speaking at the HP Asia Pacific and Japan Executive Forum in Bali, Rob Becker, HP business development manager, Business Critical Systems ANZ, said mainframes are still dogged with skills and applications problems — but replacement remains an ordeal.

With staff that wrote the original mainframe applications likely to have long since left the business, the mainframe's structure and code is hard to change. Add to the mix applications written in outdated programming languages, and it's hard enough for current IT employees to maintain the mainframe, let alone add functionality, Becker said.

The skilled workers required to maintain the machines are also becoming scarce as more mainframe professionals retire, according to Becker. "The knowledge they take with them is just not taught in [universities] today," he said.

A significant portion of the people who understand companies' mission critical systems are all eligible to retire during the next five years according to a February report by Gartner analyst Dale Vecchio, with some 25 percent to 30 percent of employees with legacy skills eligible to retire in the next three years.

"Despite efforts by IBM to establish a growing skill base for mainframes, there is little other industry-supported activity in this area. Not only is the syntax of traditional mainframe legacy languages rarely taught (COBOL, PL/1, Assembler, Software AG's Natural), but the procedural approach to application development is not the focus of modern AD architectures," he said in an earlier report.

However, replacing legacy mainframe systems is a major operation. "Taking mainframes out is like a heart transplant," HP's Becker said, as the mainframe is housing the backend functions that keep the business running. "If you tried to replace it and you got it wrong, it would cripple the business."

According to a research note by Vecchio, organisations carrying out mainframe modernisation need to consider that applications built for the mainframe over the decades differ in style from today's applications, meaning modernisation will ential an application architecture shift. Enterprises also need to take into account the fact that "the quality of service of the mainframe represents the epitome of reliability, availability and serviceability," and migrations will take a lot of effort to reproduce this quality.

HP has carried out 50 mainframe migrations in Asia Pacific and Japan over the last 18 months, according to Becker, with the Asian market holding over 1000 mainframes.

However, Australia has less than 100 mainframes, he said, which are running core systems in large companies such as banks, airlines and government agencies such as Centrelink.

These companies have been avoiding the risk of replacing the mainframe by using a surround strategy: keeping the mainframe at the core, and surrounding it with systems running applications such as net banking, Becker said. However, the strategy can't go on forever, he continued: "You can only go so far before the core itself can no longer support the business."

However, Gartner is not hailing the death of the mainframe, according to Vecchio in a separate report. "Re-platforming will be a good strategy in many cases, albeit one with significant costs and risks of its own. There will be cases where workload and QOS requirements will, at least for now, be best met on the mainframe." Companies remaining on the mainframe need to start training programs to battle the dearth of mainframe skills, Vecchio suggests.

Reaching the decision to change is difficult, HP's Becker said. The CEO and the board of the company need to be behind it; it needs a dedicated CIO, because the migration will take longer than the average CIO tenure; and resistance within the organisation will be phenomenal. "No one wants to have their heart ripped out," he said.

Suzanne Tindal travelled to Bali as a guest of HP.

Topics: Hewlett-Packard, CXO, IT Priorities

Suzanne Tindal

About Suzanne Tindal

Suzanne Tindal cut her teeth at as the site's telecommunications reporter, a role that saw her break some of the biggest stories associated with the National Broadband Network process. She then turned her attention to all matters in government and corporate ICT circles. Now she's taking on the whole gamut as news editor for the site.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • The Dinosaur is dead - long live the Dinosaur

    One of the reasons that mainframes have survived in large business are that they are still the only solution that can process the mountain of data that these business's consume. I began my career in mainframe support and eventually migrated to other platforms and am now in management. The solutions offered by Microsoft and others might be glitzy, but most of them are really struggling when it comes to meeting the throughput that you can get from an Enterprise Mainframe. Can a windows server box really process hundreds of GB of data in a timely manner?

    Throughout my long (over 30 years now) career I've heard the call that mainframe technology is a dinosaur and not the way of the future, yet today I see Windows and Linux both moving towards the type of operating environments that mainframes provided 30 years ago. IBM old catch cry 'reliability maintainability and security' is even more valid today in the world of PC & server, where functionality seems to win out over good long term strategies.

    Why are mainframes out of fashion. Because they are old and computing is still obsessed with the new rather than using the right solution for the job. Mainframe technology will be replaced when the newer technology actually is better. I seriously doubt windows server comes any where near the level of reliability that you want to risk your businesses future on it.

    It unfortunate that they don't expose students to mainframe technology early, as then they might realize that most of the new technology being implemented on servers is just a rehash of what IBM and other provided a long long time ago, but with out the graphical interface.
  • What is the definition of a Mainframe

    What makes a system able to process lots of data?? CPU, memeory, I/O??

    Sun E25K - many processors, much RAM, much I/O, virtual environments??

    Datacentre server - lots of processors, much RAM, much I/O, virtual environments?

    IBM Z series - many processors, much RAM, much I/O, virtual environments.

    Or does it boil down to the operating system?

    Or is it the processes that manage it?

    seems to me that the problems encountered in "non mainframe" systems are more related to the processes that are used to run them (or lack of them).

    capacity to process is also dependent on access to the data - which involves the storage systems as well as the box that holds the cpu.
  • One size does not fit all

    It is true that lot of organizations have been thinking about replacing the mainframes with other new technologies. But, this is a decision which is very hard to make and as rightly said by HP its like taking the heart out. If someone has to make this decision it has to be very well thought out. The reason for saying this is because it has the ability to take you out of business if something goes wrong.
    Finding out the alternative to replace the mainframes itself can be considered a project. It requires lot of effort and to find the right product to replace in itself is a ardous task. There are lot of re-platforming tools in the market which claim to simulate a mainframe like environment but cannot match the performance of a mainframe. There will be lot of patch work to be done to make the system up and running. Its a better idea to build around it rather than taking it out and replacing it with a replatforming solution.
    A replatforming solutions will work only if the technologies used on mainframes are not complex. An applications having COBOL, CICS, DB2 is probably easier to migrate rather than having something like PL1, IMS, IDMS etc. It just does not do the job.
    As far as skilled resources are concerned. It should not be difficult as there is always India around to provide that. Offshore it and be happy.