We found it very interesting that the Chinese appear to have fallen in love with the mainframe. This led us to do some research into why this should have happened. Our investigation uncovered a number of reasons, historical and cultural, that could account for this situation. Incidentally, we also learned something about the Chinese, as well as the computer industry.

The first reason for the mainframe’s successful penetration of Chinese industry, primarily in banking, is most likely historical. While, it appears that at one time, China did manufacture its own computers, even displaying one in Toronto at the Canadian National Exhibition in 1972, they were not a significant player in the general market at the time.

After the death of Mao, China began to slowly open up to the outside world. China approached both IBM and Digital Equipment Corporation (DEC) in their attempt to buy the latest computer systems. (These two dominated the market at the time. IBM had its mainframes. DEC had its mini-computers.)

As it turned out, Ken Olson, CEO and founder of Digital Equipment Corporation, was unwilling to sell DEC mini-computers to the Chinese. His avowed reason was that it was more trouble than it was worth[1]. This left a clear field for IBM. (There was a rumor that Hewlett-Packard was also interested in selling, but the Chinese were not buying[2]. In later years, HP successfully sold its servers in China.

IBM made its initial sales of mainframes into China through third party partners, only gradually expanding their corporate presence there. China initially concentrated on placing these machines into its banking industry. As a side note, societies in Asia are overwhelmingly oriented toward cash transactions. Banking is a huge business involving an enormous number of resource-intensive transactions. The mainframe was the ideal platform to handle the resulting massive workloads that had to be processed. As what is today called ‘a first mover’, IBM mainframes were able to establish an early, very solid foothold in China. Ken Olson’s policy of refusal to sell to China was not reversed until much later, and by that time, IBM was firmly established.

A second reason for the attraction and success of the mainframe in China is its underlying architecture marked by the centralization of functions and comprehensive view of, and control over, all activities. This mirrors the traditional structure of the Chinese government,   which greatly favors centralization of function and power. This held true across the early decades of the development of the computing market. The mainframe has long been viewed as the epitome of computer centralization.  Chinese business executives structure and run their businesses in ways that complement the government structure. They naturally are attracted to such a centralized solution.

Even the arrival of ‘distributed computing’ in the form of the classic three-level computing architecture off-loading Mainframes with the servers and attached terminals sold by Digital Equipment and HP during the 1980’s had little impact on the established adherence and preference for centralized computing architectures. The explosion of interest in distributed computing, experienced in the rest of the world, never fundamentally took root in China.

We should differentiate between “distributed computing” and the use of personal computers. In fact, personal computers are very widely used in China. Sometimes to the dismay[3] of the government[4], which would like to ‘see’ a little more “order” in the proliferation of blogs[5], database searches (Google) and various social applications[6]. Also, the use of personal computers (among the better educated people) has become so entrenched that there is some concern over students not leaning to properly write Chinese characters. (There are several very interesting ways to generate Chinese characters via western keyboard. The concern is that using these techniques the user does not necessarily need to remember how to completely form the character.)

As recently as January of 2012, Professor Daniel Bell writing in the NY Times about city planning in China observes that, “From the outside China often appears to be a highly centralized monolith[7].” The Professor then goes into a detailed discussion of the intense competition among China’s new cities and the differences among them that arise as a result. However, he further observes that these differences are only ones that have been allowed by the central authorities. He points out these same authorities are the ones that ultimately decide “what works and what doesn’t.” Thus, the Professor affirms that preference for central control of the planning for Chinese city development still dominates.

Another reason for the attraction of the mainframe can be attributed to the path taken during its development, evolving over time in a controlled and consistent manner. The mainframe was developed by IBM systems in the 1960’s. System/360 was announced in 1964 and it evolved over the next decade into the MVT version of OS/360.

Later, in the 1970’s, virtualization was added to the hardware and the operating system ultimately became known as MVS. IBM was careful during this period to make sure that compatibility was preserved. Thus a customer’s investment in this technology would not be jeopardized as the system continued to evolve and improve. One can see how this history would appeal to the Chinese. The technology had a history of performance and reliability that had proven itself over time. While the Chinese might be adventurous in some areas, they remain very conservative in others. One of those areas is the banking industry, which was the first industry to computerize.

We see two more somewhat connected reasons underlying the popularity of the mainframe in China. The first is the requirement for scalability. Someone once stated that even a ‘small’ bank in China has 100 million customers. The enormous size of the country’s population means that they need systems with the ability to scale to serve many millions of customers. Secondly, as mentioned earlier, the Chinese, along with most Asians, prefer to deal in cash transactions. It should be no surprises that ATMs have an enormous presence with their ability to dispense cash at all hours and at many locations. Any visitor to China will almost immediately notice that Chinese people use ATMs more frequently than an equivalent Western population. This means that transaction loads on their financial systems are far heavier than it would be for any equivalent population of customers in the West. The Chinese are convinced that only the mainframe has the ability to scale to handle both the workload size and number of concurrent users.

So now, the reasons for the Chinese love of mainframes appear to us to be relatively clear. The first platform on-site, strong centralized control of architecture and operations, evolutionary yet conservative in design and improvement, reliably able to handle massive numbers of users and transactions – all characteristics that fit the psychology and needs of the Chinese.

Some of the consequences of this situation are interesting. If you google “China and mainframes” you will find many job openings for folks with mainframe skills in mainland China or closely linked territories, such as Hong Kong. In addition, as one might expect, the IBM presence and involvement with the Chinese market continues to grow and evolve as the mainframe features continue to evolve and expand. IBM established the China Research Lab in Beijing in 1995 and added another research building in Shanghai in 2008.

Details of the number of mainframes in-place and sold to China are not available. However, secondary indicators point to the continuing popularity and growth in the number of mainframe installations there. One indicator is the close cooperation between IBM and Chinese university, enterprise and government staff as they continue to apply the mainframe of addressing the business computing challenge to flawlessly manipulate databases to handle the online transactions of 1 billion plus users.

Other indicators include investments by Chinese universities in mainframe training, IBM’s second largest Global Delivery Center focused on Mainframe competency is located in China, a Chinese driven three-year project committed IBM to train 3,000 selected graduates from six universities in Tianjin City[8] as IT mainframe and Java-based systems professionals. Simply goggling ‘Mainframe Jobs in China’ yields over 150K hits – not all of them are exclusive to China but the sheer volume is a good indicator a market for talent exists.

Finally, consistent year-over-year growth in mainframe sales in Asia/Pacific reported by IBM[9], show that the mainframe features important in China resonate in other developing countries. IBM enjoys considerable success in selling the mainframe in Korean, India, etc. In its most recent financial report[10], IBM mainframe revenues declined in quarter-to-quarter (Q4FY11 vs. Q4FY10) comparisons. It should be noted that Q4FY10 growth was the biggest in mainframe history. Earlier quarterly mainframe revenues consistently increased from the previous year. Overall, FY11 hardware revenues continued their year-over-year growth.

Any savvy IT professional in the US is well aware that much of the talk about the mainframe dying in the US market has died away. The number of new customers and users selecting mainframes as their platform of choice worldwide appears on a path of continuing growth[11].

Bill Moran

Rich Ptak

This article also appears in the May/June issue of Enterprise Executive (previously Mainframe Executive) now available here.