Not long ago, it was widely considered that, although the mighty mainframe still plays a central role in many enterprise IT architectures, it was not a technology that was growing. The fact that the mainframe market was dominated entirely by a single company – IBM – did not attract all that much attention as a result.

Recently, however, this has begun to change. In October 2009, the US Department of Justice announced an investigation into IBM’s mainframe business. The investigation relates to accusations made by a small company named T3 Technologies, which sells alternative hardware for mainframe operating systems. It says that IBM is violating its rights by preventing its operating systems from working on T3 hardware, allegations that IBM has denied.

Meanwhile, IBM’s mainframe business has come under attack from a little-known mainframe management tools vendor called NEON Enterprise Software. Having been in the business for 14 years, NEON recently made a technical breakthrough – it figured out how to run IBM software, such as database software DB2, database management tool IMS and its CICS transaction server – on mainframe hardware using so-called ‘specialty processors’.

These processors were introduced by IBM – as a response to competitive price pressure – so that certain mainframe workloads could be executed without incurring the per-cycle fees that it otherwise charges for use of its ‘general purpose’ mainframe processors. Certain workloads, that is, but not all.

NEON says its product zPRIME, launched in June 2009, bypasses that limitation. This makes using mainframes cheaper in two ways, it says. Not only does it remove the software charges associated with using ‘general purpose’ processors, but ‘specialty processors’ are themselves significantly cheaper to acquire.

“We’ve had customers telling us that this can save them 20% of their IT budget,” says NEON CEO Lacy Edwards.
“The response we’ve had has been incredible,” he adds. “We have approaching 100 companies that are in evaluation, and a number of customers that are now in production.”

Of course, this has not gone unnoticed by IBM. Around the time of zPRIME’s launch, the company circulated a letter to customers saying that, while it did not have “enough information about the referenced product [zPRIME] to comment on any specific product-related claims,” any system that allows mainframe software to be run entirely on specialty processors “should be evaluated to determine whether installation and use of such a product would violate…the IBM Customer Agreement…and/or the licence governing use of the IBM ‘Licensed Internal Code’”.

“Clearly, this is an area in which IBM is going to lose revenue,” says Edwards. “And it is reacting in a way that you would expect any business to do. But there is nothing it can do to stop our customers from using this product – we’ve checked very thoroughly.” Before launching zPRIME, he says, NEON engaged a law firm and an ‘expert witness’, both of whom said that it violates neither IBM’s intellectual property nor its licence agreements.

In fact, Edwards argues, zPRIME will revive the viability of mainframes. “Long term, the use of zPRIME will make the mainframe attractive, which it would not have been previously,” he says. Edwards argues that, in light of the cost reductions, businesses will soon be considering launching new applications on mainframes – something he hasn’t seen for many years.

According to Bryan Foss, an independent IT consultant and academic, customers are not interested in arguments between suppliers, but they are interested in cutting the cost of operations, both in the long and the short term.

Foss argues that reducing the cost of mainframe operations represents a significant opportunity to reduce expenditure in the short term, but of late it has not been the focus.

“Sometimes, organisations don’t realise how important their mainframes are,” he says. “I recently worked with an insurance company, and 70% of their policies were processed on their mainframe. But all of their new expenditure went on new applications, which were pretty peripheral to what they were actually doing.”

Perhaps that lack of focus has been a result of the perception that in mainframes nothing much is changing.
So whatever the fate of NEON, and of the investigation into IBM, such activities might well trigger a renewed period of mainframe innovation.