IT Management

Charles Dickens, contrasting different attitudes during the French Revolution in A Tale of Two Cities, wrote: “It was the best of times, it was the worst of times, … it was the spring of hope, it was the winter of despair.” The last two years have been one of the best times in the past 20 years for the mainframe despite it being a bad time for the U.S. economy and much of the global economy. Increasing sales of IBM mainframes, fueled by the introduction of the zEnterprise, indicate the mainframe’s future looks quite hopeful.

According to IBM’s review of last year, mainframe sales spiked, particularly in emerging markets—places such as India and China. This is consistent with reports from international computer analysts that, while traditional countercyclical sources of consulting revenues (such as Dubai) were unusually affected by this last recession, East Asian demand for consulting was increasing. Many of these markets have traditionally been more receptive to low, upfront cost platforms such as PCs and smartphones. Moreover, there continue to be assertions in the U.S. that the mainframe is a “dinosaur.” If emerging countries seek to establish competitive advantage by going directly to the latest technology, avoiding the burden of “legacy” systems, then the mainframe should play no part in their businesses’ IT architectures. Yet, the mainframe is clearly achieving unprecedented penetration in these markets. How can this be?

Globalization’s Evolution

The mainframe’s success can be attributed to “good technology,” and IBM’s effective marketing, sales, and support. However, until the last year or so, these factors hadn’t tipped results significantly in emerging-market sales; something else must be involved.

The “something else” starts with a shift in the way emerging economies and globalization interact. Specifically, in the old days, countries such as Japan and South Korea would start by giving local businesses advantages, and by actively seeking out the latest technologies in markets that could be commoditized. This occurred with Dynamic Random Access Memory (DRAM) technology in the ’80s. The extra cash allowed acquisitions of foreign companies and full entry of some firms, such as Sony and Samsung, into the global market. 

What seems striking about the latest set of emerging economies is that this process is now incorporating existing global companies, such as IBM, in a much more integrated fashion. India’s help desks morph into outsourcing some software development to Wipro; and China’s “Made in China” low-end products move steadily into Chinese research arms of Microsoft and IBM. The emphasis is less on “provide products here” and more on “provide services and solutions here.” Where Japan sought to provide a competitor to the mainframe, China may seek to provide a competitor to the New York Stock Exchange (NYSE).

A key to the mainframe’s attractiveness to emerging markets is that it’s a leadership platform for scale-up technology. Emerging markets are looking less at products with low up-front cost for mass distribution and more for platforms with rapid scalability and large capacity for global competition. Moreover, closer connections between companies make it more attractive to apply scale-up for an “instant global solution.” The mainframe can fully capture the rapid expansion of the emerging economy by selling to both the emerging globalizer and its global partners. This new globalization translates to a bright outlook for the mainframe in the immediate future, as old markets slowly recover and new markets continue their ascent. But what are the implications for IT buyers?

The Hub and Hoopla

Embedding the mainframe in emerging countries means a significant reduction in IT buyer risks from the aging mainframe workforce. While new mainframe uses will involve Linux on System z, the same “help desk bootstrapping” model that India has used should ensure a new generation of “z/OS administrator services from offshoring.” This solution may not be ideal, but it’s an option. Meanwhile, as the bulk of new mainframe MIPS shifts to Linux with increased demand from emerging economies, the need for arcane Job Control Language (JCL) knowledge should decrease. So, IT buyers should consider the risks of lack of mainframe expertise, but realize these risks are manageable and diminishing.

The new strength of the emerging mainframe markets leads to a bright future for mainframe scale-up and “hybrid” architectures such as the lash-up between the mainframe and scale-out blades in zEnterprise. Yes, the mainframe as a “hub” in private clouds is alive and well, and yes, pure scale-up has its place in supporting the new hot applications. However, for the most important strategic initiatives such as Business Intelligence (BI) and analytics, scale-up plus scale-out may be best in the long run.

Consider Watson. As noted in a recent blog post, we believe the hoopla about its win on Jeopardy, and the implications of that win for IBM’s analytics solutions, is overblown. Still, it does highlight an important fact about today’s markets, emerging and otherwise: BI and analytics are where the action is, and should be, for IT buyers.

But what’s the best long-term platform for analytics? As IBM noted, Big Data (massive amounts of information) and diverse data (social media, the sensor-driven Web, incorporation of data from different domains) are needed to drive a Jeopardy answer or competitive advantage. To scale for some of the big data aspects of analytics, the scale-up mainframe is often the answer for raw performance and uptime; to handle the diversity of data types, the combination of scale-out clustering and networking—plus a certain degree of integration that blade architectures provide—dictates traditional grid-type computing and provides a degree of “asymmetrical parallelism” or “specialized processing” appropriate to handling multiple data types.

An architecture that combines the flexibility of scale-out and the performance of scale-up may be the best of both worlds. It warrants IT buyers kicking the tires of the hybrid architecture to see if it’s the best fit for today’s needs (a core analytics platform).

Conclusion

The mainframe’s delay in Windows platform support may eventually place an upward limit on its reascent. Still, the mainframe’s future looks infinitely brighter than it did about five years ago, when IBM first started effectively placing it in the same software universe as the rest of the computing industry. For mainframers and IT buyers alike, this is a spring of hope.