We tend to get very excited about births and deaths, not just human or pet, but in the technology market. We are currently talking about the death of the PC and the “post-PC era,” but a few decades ago, we were talking about the death of the mainframe and that Client Server (read: Sun) would take it out. Well, over three decades after we pronounced the mainframe dead and Sun the winner, Sun no longer exists and the mainframe is doing pretty damn well.
BMC just finished its 7th annual mainframe survey and while this market isn’t growing appreciably, it isn’t shrinking either; in fact, according to the survey, the very largest of companies are putting more and more of their mission-critical loads on them. On my last check, this class of system remains IBM’s most profitable and IBM, financially, is currently towards the top of its historic performance.
Let’s explore the survey results.
I did much of my undergraduate work and graduate work in market research and by profession I’m an analyst. The survey methodology uses a fixed sample of self-selecting respondents, which means it isn’t a survey that can be used to talk about the broad IT market. However, because it had 1,264 respondents and the sample has been kept highly consistent, it does reflect what is going on in the mainframe shops sampled (this was an extensive survey of shops that currently use mainframes).
This means we can specifically look to see if there is an emerging trend to eliminate mainframes, but this survey would not capture any trend for non-mainframe shops to acquire them because these shops would not be sampled. This means the results should be biased negatively with regard to market share growth for this category.
The makeup of the sample was predominantly U.S. (60 percent) with EMEA (30 percent) and Asia-Pacific (10 percent). Respondents were non-executive practitioners (77 percent mostly and 17 percent were executives). Company size was slightly weighted towards those over $1B (54 percent), with 43 percent in the Financial and Insurance markets (where mainframes’ ability to handle massive datasets assure they remain very common) and 19 percent government.
The top IT concerns, however, appear to map well to non-mainframe shops with some interesting changes over the last year. The highest priority remains reducing IT cost, this is followed by disaster recovery, showcasing we continue to exist in hostile times. Next is application modernization, which continues to showcase the changing IT landscape for technology solutions.
A big change this year is Business/IT alignment, which jumped from 7th to 4th, likely reflecting the growth spurt and the need for IT and line management to be in sync. Dropping from 4th to 5th is MIPS efficiency, which is interesting given the massive focus on the IT industry on exactly this topic as a competitive advantage. Reducing the impact of planned or unplanned outages dropped from 4th to 6th, likely showcasing an increase in relative reliability resulting from the prior focus; in short, many addressed what had been a problem successfully.
This seems to map well with other surveys I’ve seen over the last year and suggests a sample that, at least with respect to priorities, appears to be relatively generic. If you have what appears to be a non-scientific sample, you can do a mapping to other samples giving it more apparent weight.
This would explain the results that suggest that 6 percent of respondents were exploring getting rid of their aging mainframes, but this is a far cry from the a similar survey I recall from the 1980s, which suggested the majority of mainframe users were going to exit and clearly that didn’t happen. Four percent of respondents were planning on doing things with their mainframes that fell outside of the survey’s ability to capture.
Of the remaining 90 percent majority, 50 percent were planning to put more traditional workloads on the mainframe and 40 percent planned to use the mainframe for new apps.
So while it is hard to argue mainframes are growing, it is equally hard, in this decline biases survey, to argue they are declining and it is clear they are being moved to contemporary loads and tasks aggressively.
The reasons they are preferred remains relatively constant over the last 3 years. The unique nature of the platform, security and centralized nature are the top three. It remains a successful specialized server for high transaction loads and in that area, to these respondents, it remains their favored choice.
It does appear that, regionally, Asia-Pacific is planning to increase its spending on mainframe software significantly ahead of the rest of the market and that the market in general is becoming far more aggressive in using mainframes for new applications with a significant emphasis on service-oriented architectures and Web services.
Training has shifted strongly towards internal methods supplemented by validated experienced external candidates. Outsourcing is a much smaller number and if the moves at GM are reflective of the market, outsourcing is losing steam.
This annual BMC survey indicates that not only is the mainframe alive and well, it is evolving to take on newer loads and, as validated by IBM’s financials, is doing better than some of the newer architectures (where is Sun again?). In the end, this survey just stands to remind us that we really don’t change that quickly or easily and sometimes the old tried-and-true methods do work better, particularly if they can be advanced to embrace new concepts and ideas. For major initiatives like cloud computing and software as a service, the transactional advantages that mainframes have may keep them around longer than we are.