Mainframes are an extraordinary story of technology survival, having lasted from the very beginning of the computing era and survived PC-based computing, Y2K code rewrites and the onset of cloud. But what would drive someone to buy a new mainframe today?
Picture: /Getty Images
The existing base of mainframe users in Australia is relatively small, probably under 100 sites. But that market remains lucrative enough for upgrades to continue. Last month, IBM rolled out its new zEnterprise BC12 mainframe, and upgraded its mainframe operating system z/OS to version 2.1.
"It's a well rounded product that is based on a long history and is highly scalable and very affordable," Andreas Wenzel, business unit executive System Z for IBM's ANZ systems and technology group, told Lifehacker.
Prior to working in the ANZ role, Wenzel headed up the IBM's global system enterprise initiative, which was "solely focused on selling mainframes into clients who had never heard of mainframe before". According to Wenzel, more than 200 customers signed up as new IBM mainframe users last year.
One factor in those sales is that mainframes can now be used to run Linux environments, which offers access to apps not available in the traditional z/OS environment. One flavour of the new model, the Enterprise Linux Server, sells on a three-year support contract and includes the hardware and a Linux hypervisor.
Another familiar driver for mainframe usage is the increased interest in cloud computing. Mainframe systems introduced one of the core cloud principles: you pay based on usage of specific apps. "If you look at the fundamentals of cloud — virtualisation — that was always there in the mainframe," Wenzel said. "With cloud becoming more and more prevalent, I think there's a strong argument for the mainframe to be a way to get into this cloud model easier, faster, cheaper and better."
Mainframes also have potential in the area of big data analysis, and can eliminate the latency challenges associated with cloud-based big data systems, Wenzel said. "You have a closeness to the data, and you have a transaction engine where you can run all of that. It reduces latency in the network space and gives you more responsive outcomes. When someone uses a mobile device to initiate a transaction, you can actually drive real-time analytics as your customers handle or order your product, rather than having it in a distributed environment."
Mainframes don't come cheap: a financing deal for one will cost you at least $US1965 a month, or you can buy one of your own from a starting price of $US75,000. Surprisingly, some customers still prefer to buy outright. "I would always tend to recommend going down a financing pat, but in some areas they just want to spend the money," Wenzel said.