The IBM mainframe has been a stalwart platform for enterprise computing for the last 50 years. There are not many technologies you can point to that have survived that long and remain a key component of many companies’ IT infrastructure. Its survival should be lauded, but the question remains, is it the only game in town for enterprise computing. That answer is a resounding no. The mainframe is a combination of hardware and software that is arguably unique to all other computing offerings available in the marketplace today. That doesn’t make it a bad platform, but it does limit its opportunities for growth. It is a proprietary platform, only available from a single vendor – IBM.
More importantly, it is dependent on a skill set that is most commonly found amongst the Baby Boomer generation – a generation that represents the largest retirement demographic in the history of the galaxy! Younger generations could clearly learn the technologies of this platform, but they haven’t. There is little proof, in spite of efforts by IBM and others, to grow a workforce of the scale necessary to replace mainframe skills around the world. IBM has invested in this platform for decades, although recently it seems more focused on selling off mainframe software products and the development and support staff associated with them. The future for IBM, in its own words, has become about cloud and Watson. You rarely hear about the mainframe from IBM when it is discussing the future. IBM didn’t even build Watson on the mainframe, it was built on x86 infrastructure.
The overwhelming power of x86 environments and the resiliency and performance being built into cloud infrastructure platforms is unparalleled. The reality of delivering large-scale, enterprise-class workloads in these Infrastructure-as-a-Service (IaaS) environments is here today. This fact doesn’t mean the mainframe is a bad platform. It only means there are more cost effective, agile and innovative environments that can provide high quality infrastructure for running enterprise workload. The challenges are not whether these modern platforms can handle the workload – Amazon, Google, Ebay, FaceBook and others have proven that. The question is whether organisations are willing to take the risks in modernising their application portfolio and what they are willing to spend to do it. During my 18 years as an analyst at Gartner, I took more mainframe inquiries than anyone there. Clients WANT to get off – they have shown reluctance because of the risks of moving. Also, the defenders of the platform have always thrown up roadblocks to migration since it threatened their power and even jobs. An understandable reaction by the individual, but an organisation’s IT strategy shouldn’t be driven by the retirement plans of their employees!
Tale of Three Cities
The IBM mainframe market is really a Tale of Three Cities – small, medium and large. Small organisations are willing to get off, can get off without great risk and have the greatest ROI since they pay more for IBM MIPS than anyone else. Large mainframe shops are inclined to move some workload off the platform to commodity x86 environments even if they aren’t willing to commit to a complete migration. The challenge is for medium sized mainframe shops to strategically accept the migration even if the first few thousand MIPS that are moved don’t generate significant cost savings. They have to stay the course, or they will give up before real benefits are achieved.
As Baby Boomers continue to march inexorably towards retirement and the realities of cloud environments is accepted more mainframe workload will inevitably move to this world. Many organisations don’t want to be in the “IT business” and are much more accepting of a cloud provider supplying the equivalent of IT dialtone. Workload will continue to run within in-house datacentres, but that will move to x86 multi-core environments as well. The IBM mainframe had one strong run, but the power of modern computing is x86. Is the mainframe dead? No. Will it become increasingly marginalised to the high end of computing? Most likely.
Dale Vecchio, Chief Marketing Officer, LzLabs
Image Credit: Majestic B / Shutterstock