The rumors of the demise of the mainframe are greatly exaggerated

The rumors of the demise of the mainframe are greatly exaggerated

Summary: Although suppliers of midrange machines would have us believe that the mainframe is dead, IBM's System z continues to grow.

TOPICS: Dell, Hardware, Servers

IBM' System and Technology Group (STG) offered analysts an update on the success of its System z family of mainframes. I wasn't at all surprised to learn that current users are buying more of these systems and new customers are buying them.

What IBM had to say

IBM presented a list of impressive statistics designed to prove that the mainframe is alive, continuing to serve in its original mission of being an organization's central computing system, support for massive number of business transactions or platform for very large, centralized databases. It is also claiming new workloads because of its tightly integrated management and security capabilities.

Here are a few tidbits from IBM's presentation:

  • Total installed System z capacity continues to grow.
  • The centralized, unified management that is offered by IBM zEnterprise Unified Resource Manager, is attracting new customers. Unified Resource Manager makes it possible for a System z configuration to manage blade systems running AIX, Windows or Linux
  • IBM continues to innovate in the areas of system to system communication allowing hybrid mainframe/UNIX/Windows/Linux environments to provide high levels of performance. This allows organiztions to re-centralize distributed applications to reduce overall costs.
  • IBM's zEnterprise system has been purchased by over 120 new customers, customers from all over the world and in IBM's target markets


Time and again, I hear suppliers of midrange systems, such as Cisco, Dell, or Oracle/Sun, present the story that their systems can address any computing problem from the very small to the very large. Although it is often unspoken today, their message to big organizations is that they can discard their mainframe installations and move those workloads onto clusters or grids of midrange systems without worry.

Why is it, then, that IBM's System z family continues to grow? Why do mainframe configurations find their way into the data centers of new customers? The answer appears to be the same now as it was back in the 1970s and 1980s.

For some workloads, a single, very large system configuration works better and is easier to manage that a veritable herd of industry standard systems. If one examines cost of ownership or return on investment studies, staff-related costs of administration and operations far outweigh the cost of systems and software. So, any platform that minimizes those costs, can actually lower overall costs.

What IBM has done is to make the mainframe a hub of computing that can support and manage applications designed for mainframes, for UNIX, for Windows and for Linux in a unified and secure way. IBM also provides tools that allow highly distributed systems to be managed in a more unified way as well.

If anything, IBM faces the challenge of having many good solutions that can be used individually or together. This complex product portfolio can be quite daunting for some.

The key challenge IBM faces, of course, is getting IT decision makers who grew up using industry standard systems and thinking that mainframes were holdovers from some per-historic time to actually become aware of what these systems can do today, compare their own company's costs with and without the mainframe and come to the decision that today's mainframe is a valuable addition to the data center, not a relic of the past. IBM's Smarter Planet marketing campaign is designed to do just that.

Can IBM succeed in re-introducing mainframes into an industry standard world? The facts appear to say yes.

Topics: Dell, Hardware, Servers


Daniel Kusnetzky, a reformed software engineer and product manager, founded Kusnetzky Group LLC in 2006. He's literally written the book on virtualization and often comments on cloud computing, mobility and systems software. In his spare time, he's also the managing partner of Lux Sonus LLC, an investment firm.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Should I be singing and dancing etc?

  • Is it really the same?

    It seems like to me that they are just repurposing the mainframe as a VM host for Unix/Linux.
    • Is that a problem?

      Centralized management can reduce the cost of administration. This, in turn, is a major cost factor for application systems.

      So, is it a problem if the IBM folks have developed ways to manage and host mainframe applications, UNIX, Linux and even Windows?
  • Margins are incredible

    IBM's profit margins on mainframes are incredible. mainframes are not going anywhere soon. Nothing today quite matches the horsepower of a mainframe, despite the power of midrange servers.
    Your Non Advocate
  • Energizer bunnies

    Another issue for many customers is uptime. Mainframes regularly deliver 99.7% and even 99.9% uptime. For a bank, and for many other industries, the difference between 99.8 and 99.9 is worth millions of dollars.
    Robert Hahn
    • RE: Energizer bunnies

      It's easy to forget what takes to accomplish that uptime. In my days as a console operator (circa 1996), we had a whole room of console operators responding to factors that lead to system degradation in order to keep the system from needing an IPL (or reboot). Maybe things are better now.
    • That's just the beginning

      The proper system configuration that includes redundant processors, memory, storage, network interfaces combined with the proper software technology can push that number up a couple more nines.

      This can mean the difference between 8.76 hours of downtime a year when a system provides 99.9% uptime and 5 minutes of downtime a year when a system provides 99.99% uptime.
  • design vs cost vs relevancy

    I spent 20-years working for IBM in mainframe new technology, and am now at Dell as Director of Systems Engineering. What IBM told you is totally true, more installed MIPS, management of mixed environments, etc. etc.

    Context is everything though. Total mainframe MIPS will likely continue to grow inline with their ability to keep pace with general improved price performance. Customers will move to other platforms based on the pain of IBM hardware and software pricing. Thus if IBM are doing the right things installed MIPS will grow because customers are not moving and their workloads are growing, more data, broader customer sets, web driving higher transaction volumes.

    Add to that customers have set buying cycles, refresh cycles etc. and each time IBM comes out with a new processor, it has more MIPS/CORES etc. Even if they may not losing customers, growth in MIPS doesn't mean they are gaining them.

    Now, the question to ask is, is the growth of mainframe MIPS inline with the growth of non-mainfarme MIPS?

    in the x86 world, we are still making hardware design decisions based on cost, where IBM mainframes have already solved the problems and paid the price many years ago. The question is does it make sense to chase those final few pieces of design, when software is increasingly becoming abstract and in the "cloud" where the software replaces those hardware features with their own redundancy, high availability etc.
    • In a word, yes.

      First of all, thanks for stopping by and commenting!

      Software-based failover mechanisms may be able to keep applications running. The failover time, however, may not be short enough to prevent application slow downs or failures. A failover in minutes or hours may not be acceptable when the application requires failover times in microseconds.
  • Dead, demise thats just lazy anyway

    oh yeah, from my time at IBM defending the mainframe, people that say anything is dead are being lazy, sloppy, it's just the best way to start a discussion about why you should do something else...

    See my blog on the topic here:
  • Better Late then Never!

    Dan, Paula, & Ken,

    You are 2.5 years and half a trillion dollars late!

    What took you so long?