Advertisment

Mainframe not yet a has-been: IBM

author-image
CIOL Bureau
Updated On
New Update

SreenathGeriatric, bygone, past one’s prime, even, anachronistic; one may be too inclined to attach these adjectives with some technologies. But when it comes to Mainframes, IBM strongly contends that they are not obsolete yet.

Advertisment

Incidentally, in a recent survey by CA reliability, resilience, greenness, affordability of staff, and security have come up as the upsides and compelling advantages of the so-tagged out-of-mode mainframes as enterprise platforms.

The survey of over 100 IT executives not only highlighted the persistent appeal of the mainframe as a cornerstone platform for multi-tier enterprise computing, but also pointed out how cost issues and limited data centre space tip the scales from distributed platforms in form of Mainframes.

An interview with Sreenath Chary, country manager – System z, STG – IBM India/SA tells us more.

Advertisment

How do you assess the relevance and survival of Mainframes in the context of emerging disruptive technologies , Unix, Windows NT, cloud models, green data centres, storage dynamics?

For decades, the Mainframe was inarguably the leading data processing machine in the world's largest corporations. Some believed it would eventually become obsolete, but IBM continues to prove them wrong. Today, most of the world’s largest enterprises rely on Mainframes for their mission-critical transaction processing and data serving.

About 80 per cent of corporate data resides or originates on the Mainframe due to its superior security, virtualization and collaborative processing capabilities. IBM has gained market share back to what it was back in 1992, which was the pinnacle of its success. System z is extremely relevant as a number of these technologies took birth in the Mainframes.

Advertisment

Hasn’t the realm of Mainframes suffered with the challenge of skill obsolescence?

System z with zOS does have an issue with graying of the skills, but z is possibly the most open platform today – it runs zOS, zVM, Linux, zTPF, DOS/VSE operating systems natively. No other platform can do this. The fact that it runs Linux natively allows for any Linux skill to exploit System z. Hence the statement of skills obsolescence is only in a portion of the total dynamic.

What kind of impact has happened in particular with virtualization and cloud computing?

Advertisment

Virtualization started with z (zVM) and is possibly the most mature of virtualization technologies today. System z Virtualization is surpassed in its capability for sharing memory, I/O and processing. It emulates entire networks in memory making investments in switches, routers and networking gear redundant to a large extent.

As to Clouds - given that the Mainframe is the best environment to run mixed workloads with reliability and assured throughput, z also forms the best to implement cloud computing, as it can deliver technology like capacity on demand, deliver virtual servers on demand.

This technology for capacity on demand also originates in System z and is now catching on Green data centres - System z is the cornerstone of our new enterprise data centre. IBM has consolidated 3900 servers into 30 System z servers saving over 85 per cent in power and over 80 per cent in data centre space.

Advertisment

Coming back to the skills-related challenges, how is IBM taking it on?

Linux skills are available in plenty. We established the Mainframe Center of Excellence to enable enhancing mainframe skills and are addressing the skills issue for zOS through our academic initiative.

IBM is taking the lead in developing the next generation of mainframe engineers in India by launching many initiatives, one of which is the IBM Academic Initiative: System z Program. This is a joint program with 230 universities worldwide that is designed to train 20,000 mainframe engineers by 2010.

Advertisment

It provides colleges and universities with educational resources to enhance the IBM System z skilled resource base and help students develop practical skills that enable them to find good jobs quickly upon graduation.

How do you reckon the increased migration of mainframe legacy environments to new ones. Is Hitachi's and Amdahl's exit an indicator of the change in scenario?

Mainframes are a legacy for us and we will continue to bring breakthrough innovations in mainframe technology. Through the years we have made mainframes available to all sectors and verticals and have reached a level where we can bring a technology like the z10 to the market which is rated as one of the best so far in terms of its breakthrough technology.

Advertisment

IBM spends close to $1billion in research and development around the Mainframe. Innovation happens on the mainframe first.

Exit by Fujitsu and Amdahl has been, in my opinion, more from the fact that they could not provide innovation to the same level rather than being an indicator of a technology in decline.

Are software costs and staffing issues making the Mainframe continuity difficult for CIOs?

Yes, we do have some small challenges in the lower end of the market (less than 500 MIPS), but the same is not true for larger accounts.

IBM Software and the zStack becomes a lot less expensive and can be shown to be the lowest in costs of operation and ownership with scale. This is an environment where the unit cost of computing drops with scale rather than increase with scale (as in distributed environments). With the introduction of the offload specialty engines for Java (zAAP) and DB2 workloads (zIIP) this cost dynamic is making it a lot more attractive for customers on pricing.

What has the change in landscape been like from COBOL era to OLTP etc and how has it redefined Mainframes?

There are an estimated five trillion lines of COBOL code that currently runs. A lot of the code is for OLTP programming and the rest for batch applications. It is going to be a gargantuan task to rewrite this into other languages. While some movement to rewrite OLTP applications to JAVA has happened, the majority of the COBOL code continues to be run.

What impact are concepts like cluster computing, grid computing, massive processing making on this technology?

As discussed, System Z is extremely relevant as a number of these technologies took birth in the Mainframes. Cluster computing (delivered seamlessly using virtual servers in zVM), grid computing (utilizing unused capacity from the grid) are similar to cloud computing which can be very elegantly implemented on Z. There is no challenge that Z will not be able to deliver these technologies.

Any examples on the enterprise level, that you can site?

HDFC Bank is a perfect example of IBM's commitment to help organizations build streamlined, more dynamic infrastructures. This win is significant as it bears testimony to the value of the platform through its adoption by a perceived technology leader in the banking space. This can expand, as HDFC Bank, as a growing bank, addresses other business requirements or refreshes other mission critical applications. Further this sets the stage for many more organizations to follow suit.

So you maintain that Mainframes are relevant enough for enterprises?

Banks and other financial services companies are facing increasing business pressures - managing risk, governance and compliance issues, being prepared for cyber attacks, and managing cross-border transactions, to name a few. There is a need to securely manage hundreds of millions of concurrent transactions as their business grows - to not only serve customers, but identify patterns and trends in managing risk.

Given these requirements, a range of financial services companies are building their infrastructure using IBM System z servers. HDFC Bank is among more than 150 clients from multiple industries that turn to IBM's System z to help increase productivity and efficiency.

These banks are seeing benefits that include service improvements from "smarter" technology investments such as IBM's more energy efficient technologies that take up less space, keep energy costs low, and require fewer IT gurus to operate. A single z10 EC server is the equivalent of 1500 x86 servers, up to 85 per cent less energy and 85 per cent less data center space.

And what about security and consolidation issues?

In a consolidation, the System z10 BC may provide up to six times the same work in the same space and may provide up to 16 times the work for the same power consumption - given its virtualization capabilities - to simultaneously manage a variety of business operations.

Increased security capabilities to run a highly responsive and business-driven IT engine that also helps meet regulatory compliance and litigation avoidance - as the industry's only commercial server with EAL5 security classification.

The ability to securely manage hundreds of millions of concurrent transactions stays. In fact, in 2007, benchmark tests with Bank of China confirmed a record 9,445 business transactions per second (tps) in real-time based on more than 380 million accounts with three billion transaction histories.