Finally, the last and most common theme for the slow adoption of the Penguin was that running Linux on the mainframe is an exception. In many respects Linux on System z is the same as running Linux on other platforms such as x86; it’s just a different kernel. Or at least that’s what Linux on System z vendors like to tell themselves—and you.
In reality, there are some pretty big differences to running Linux on System z rather than on a distributed box. For one thing, there are a lot of required tools: tools to manage the hypervisor and the underlying hardware, plus tools for system-related functions such as monitoring, performance and resource utilization, backup and instance provisioning and management. Each hypervisor requires its own unique tools, and what works on the distributed platform most likely won’t work for z/VM guests. Therein lies the exception. Not only do enterprises have to purchase additional tools to manage the Penguin on System z, they also have to teach their staff how to use them.
Some of these themes can be more easily addressed than others. Focusing on the last three is likely to provide the biggest impact on helping convince organizations that Linux on System z is mature and definitely proven.
While it would be impossible to provide mainframe education to all distributed teams in companies that own mainframes, it is possible to provide tools and information to mainframe advocates on how to successfully start and finish a Linux on System z proof of concept. Mainframe advocates should seek guidance to help them engage their enterprises and complete proofs of concept. They should also leverage forums and papers to learn tips from those who have successfully adopted Linux on System z at their enterprises, as well as how to avoid obstacles and potential pitfalls from those who had unsuccessful forays into Linux on System z.
For some organizations, a $200K hit when reaching the limit of their current IFL is not easily justified. They consider off-platform solutions instead of paying this steep cost. The suppliers and vendors that enable Linux on System z need to address this cost issue, not necessarily by lowering the pricing but by making the pricing more digestible. It’s much easier for a manager to ask for budget in $10K blocks than $200K blocks. Some managers may even be able to spend $10K without higher approval, but that is usually not the case for $200K expenditures.
Finally, and most important, IT vendors need to stop treating the Penguin on System z as an exception. There is a common belief that extending distributed tools to the mainframe taints the purity of z/VM and the mainframe. Yes, there is a place for specialty tools to manage the care and feeding of the z/VM environment, but higher-level management applications need to extend across the enterprise. If IT vendors and enterprise mainframe teams want the relationship between the Penguin and the mainframe to thrive and grow, distributed tools need to be extended to the mainframe. Using the same tools across the enterprise not only eliminates objections to mainframe Linux but also removes cost obstacles. Otherwise, the Penguin and mainframe advocates will continue to run into obstacles and challenges from their distributed counterparts. This is not a magic bullet and does not guarantee cooperation and/or success, but it will go a long way toward success.
Let’s Get Married!
It’s been almost 15 years! It’s time to cement this relationship and propel Linux on System z to success in as many enterprises as possible. Mainframe professionals, engage your distributed counterparts and challenge your vendors to enable your enterprise tools on the mainframe. Mainframe IT vendors, engage with your customers and learn how your solutions and tools can connect them with their distributed counterparts. With support and, more important, action, the success and marriage between the mainframe and the Penguin can be assured.