Escalating cost, inability to support business agility and the limited availability of modern software are some of the key factors driving mainframe customers to migrate from the age-old platform, according to HP executives.
Herbert Zwenger, Asia-Pacific and Japan vice president and general manager of business critical systems, said mainframe customers spent a large portion of their datacentre budget in licensing, supporting and maintaining a mainframe environment.
Speaking at a conference in Tokyo on Thursday to champion the company's "mainframe alternative" strategy, Zwenger said HP has helped such customers cut operating costs by up to 70 percent and reduce energy consumption by up to 46 percent, by migrating to the vendor's Integrity servers.
HP's bid for the mainframe market stems from chief executive Mark Hurd, who has never been coy about the company's plan to target and convince mainframe customers to migrate, having once said that he was "not a big fan of mainframes".
Satoshi Kitamoto, Japan head of mainframe alternative business at HP's technology solutions group, noted that mainframes were first introduced two to three decades ago as the best platform to support mission-critical processes. Today, however, users can no longer accelerate their business effectively because of mainframes, Kitamoto said.
The advent of the internet has changed the way businesses operate, he noted, and companies today must have the flexibility and agility to react quickly to competitive pressures. The ability to do so is proving to be a challenge for mainframe customers, he said.
Kang Won Moo, HP's Asia-Pacific and Japan sales director for business critical systems, added that mainframe applications and software upgrades are high and costly to maintain.
In addition, new applications developed by independent software vendors (ISVs) are mostly made available on industry-standard platforms, not mainframes, Kang said. As a result, proprietary mainframe users have limited access to modern software, he said.
Still alive and kicking
In a previous report, however, IBM told ZDNet Asia that it is seeing a re-emergence of mainframes in the region as companies look for ways to increase space and power efficiency in their datacentres. The company also noted that more ISVs now support their applications on the platform.
In fact, IBM said it invested tens of millions of dollars over the past couple of years on programmes to train customers, tweak their software and help companies migrate to the platform.
IBM has been promoting its System z mainframes and speciality engines such as Integrated Facility for Linux since the end of 2005, Erica Gadjuli, Gartner's principal research analyst for server markets, said in an email interview.
Such efforts from mainframe vendors, Gadjuli said, were the primary reason why revenues for the Asia-Pacific mainframe market, which has always been niche and project-driven, increased by 42 percent in 2007.
One of the remaining dominant mainframe providers, IBM continues to introduce new features and functions to address legacy and new workloads, added Jennifer Wu, research vice president of server markets at Gartner.
"Many of the new features are aimed at improving the ability to handle workloads that had not been traditional strengths of the mainframe, such as Java, business intelligence and datawarehousing," Wu told ZDNet Asia in an email.
Gadjuli noted that the banking and public sectors were the two mainstay verticals driving the Asia-Pacific mainframe market in the past couple of years.
However, growth for the mainframe market probably slowed in 2008 due to budget delays, revision on major spending during the Beijing Olympics and the credit crisis, she said.
Gartner expects the annual growth rate for the mainframe market to remain flat over the next five years, as customer deals typically require longer lead times, planning and execution, she noted, adding that lower pricing strategy will also drive revenues down significantly.
Now more affordable
Once seen as the Rolls Royce of computing platforms, mainframe pricing structures were previously rigid and based on assumptions that high ceilings of capacity would be met, said Madan Sheina, principal analyst of software applications group at Ovum.
"That meant companies often over paid for large CPU capacity, even though it was in excess of their requirements," Sheina said in an e-mail interview. "But that has changed following the introduction of new, granular 'sub-capacity' pricing models that allow customers to pay only for the mainframe CPU capacity they intended to use."
"Sub-capacity pricing reflects the industry's movement toward server consolidation and sub-divides the mainframe's core CPU capacity into logical partitions. It also provides a cost-effective way of adding new BI workloads and applications to run on the mainframe."
This has helped 'recast' mainframes as a more affordable platform for business applications such as operational BI, the analyst said.
A resurgence of interest in mainframes is also driven by current trends toward server consolidation, virtualisation and green IT infrastructures, Sheina said.
"Mainframes have sizeable capacities and are natural IT consolidation platforms," he explained. "They are also energy- and space-efficient platforms that, if properly configured, can provide the same processor capacity of a distributed server farm while using just a fraction of the amount of electrical, HVAC cooling and floor-space requirements, compared to a distributed server farm."
In addition, mainframes now incorporate richer functionalities such as specialised and parallel processors...
