IBM announced yesterday that they are doubling down on their Linux on the Mainframe strategy. They have embraced a more open architecture and announced LinuxONE: a hardware, software and services initiative to make the platform more accessible. Ultimately, this announcement opens the door for IT to deploy Linux on the mainframe in a much more standard way supporting KVM, using more common open source components, OpenStack management tools and removing the need to invest in the exorbitant initial outlay of a mainframe.
So, what will this change?
There are two things that no one can dispute – Linux usage is growing off the charts and that the mainframe is a monster of a platform. Linux was first ported to the mainframe 15 years ago and yet it has never really taken the share of the platform market that it deserves.
To understand why adoption has been slower than you’d expect, consider some basic steps involved in making the decision to deploy Linux to the mainframe…
| Buy a mainframe |
=> |
How much? |
| Learn all about z/VM |
=> |
z/V…what? |
| Deploy Linux |
=> |
OK, I got that bit |
| Develop customer code |
=> |
From scratch? Really? |
Most of these steps were exceptions to the “standard” process when deploying onto mainframe that IT performs or they represent barriers to adoption. Most IT shops are no familiar with the mainframe hardware, z/VM and most open source software is not ported to the platform.
Now consider the brave new world post-LinuxONE:
| Pay as you go for a mainframe | => |
Like the cloud…ok! |
| Use KVM |
=> |
Got that covered already! |
| Deploy Linux |
=> |
OK, I got that bit |
| Utilize open source components to build apps |
=> |
Yup, that’s covered too! |
As you can see, in this new model both the deployment process and the infrastructure are very similar to those that IT uses today. They can even use Docker containers to deploy applications. IBM will be detailing the elastic pricing in the near future but it will allow for the hardware to be leased with monthly payments and the software made available on some kind of usage-based fee structure removing the need for a customer to have a huge initial outlay to buy into the platform.
Now the way is clear to allow corporate IT to consider the benefits of z as a platform without worrying about the reality of whether it would actually work! In my next posts I’ll talk more about the new Special Interest Groups for Open Mainframe and the actual benefits of z as a platform.