A recent survey by CA (formerly Computer Associates) highlights the ongoing appeal of the mainframe. For example Coach Factory Outlet Online, 76 percent of respondents complained that managing a large number of distributed servers has become a cost issue. That’s a red flag.
For years, industry pundits and reporters have repeatedly warned that the mainframe was a dinosaur marching inexorably toward extinction. Yet the mainframe remains because IBM did a few smart things with it. For instance, it took advantage of the most robust virtualization engine in the industry to enable the mainframe to run virtual Linux servers. Now companies can run Linux applications on their mainframes along with everything else already running there. It also can play the SOA, Web 2.0, and mobile computing games. In short, it taught the mainframe the latest dances. For more, check out my blog, dancingdinosaur — mainframe computing in the 21st century
So if the IT guys want to ditch the mainframe in favor of sexy new commodity systems because the mainframe is too expensive, send them back to figure out how much those commodity systems will really cost over the next 3 to 5 years. And don’t let them skip the fully loaded people, energy, and shared overhead costs. ###
The issue with the mainframe really revolves around cost. The mainframe suffers from a high cost of acquisition compared to commodity Linux and Windows systems. So, when your IT guys argue that commodity servers are cheaper than the mainframe, they are talking about the acquisition and immediate deployment costs only.
But even there, IBM has smartened up and figured out numerous ways to reduce costs without outright cutting prices. The Salt River Project, a major utility in Arizona, actually shifted workloads from distributed UNIX systems to the mainframe running Linux because IBM offered key components for free. Hard to beat free. You can see the full case study here. You can get offers like that or better.
A posting on LinkedIn of a Network World piece described the Union Pacific Railroad’s long, painful, and still incomplete transition from mainframe-based computing to something new and sexy and maybe better and cheaper. It generated considerable discussion around whether the mainframe was a dinosaur verging on extinction.
The true advantage of the mainframe only becomes apparent when you look at the full cost of ownership over time. Computer hardware and software are cheap compared to the cost of people and energy. You can run far higher workloads with far fewer people and far less energy on the mainframe than on commodity platforms, even using the latest virtualization and blade technology.
Maybe your IT people say the same thing about the organization’s mainframe, usually in the context of wanting to get some slick new IT platform, something using Linux or blades and capable of SOA and Web 2.0 and mobile instead. Their basic argument is this: For what we spend on the mainframe, we could buy a ton of this other hot stuff and have money to spare.
You can’t really know which platform, mainframe, or commodity systems, is cheaper until you do a full total cost-of-ownership analysis including all the costs of people, energy, and shared overhead. You can see a fuller discussion of this here.
Related：
没有评论:
发表评论