IBM Advances Mainframe Container Strategy
IBM, as part of a larger strategy to make it possible to run cloud-native applications on mainframes, this week announced IBM z/OS Container Extensions, which makes it possible to run applications built using Docker containers on the z/OS operating system.
Barry Baker, vice president of software for IBM Z, says IBM also plans to make available other critical cloud-native technologies such as Kubernetes available on z/OS, in addition to the multiple instances of Linux that IBM already makes available on mainframes.
Baker also notes that IBM expects Kubernetes for z/OS and other capabilities will be one of the primary opportunities for IBM and Red Hat to work more closely together once IBM’s acquisition of Red Hat is closed.
One of the primary benefits of this approach is that IBM will be able to bring the quality-of-service capabilities of z/OS operating system to cloud-native workloads deployed as a private cloud either on-premises or, one day soon, on a multitenant public cloud service from IBM, he adds.
As part of that effort, IBM launched IBM z/OS Cloud Broker, an instance of the Open Service Broker Application Programming Interface (API) specification, an open source project that advances hybrid cloud computing by making it easier for developers to discover and invoke services residing on other platforms. That capability will play an instrumental role in advancing IBM’s goal to foster the creation of hybrid cloud computing environments spanning public and private clouds running on IBM mainframes, says Baker.
IBM this week also announced Tailor Fit Pricing for IBM Z, which is an effort to make available a true consumption-based pricing option. Historically, the cost of running an IBM mainframe was based on a complicated four-hour rolling average of consumption of workloads. That approach allowed some savvier IBM customers to essentially run some workloads for free on an IBM mainframe. Baker says IBM is now trying to bring more transparency to its pricing model without necessarily raising customer costs. IBM plans to work with customers to determine which pricing model makes the most economic sense for them given the characteristics of their workloads. But Baker admits IBM would like to move customers away from the existing four-hour rolling average workload pricing model whenever possible.
IBM is making clear that even after five decades of use by enterprise IT environments, mainframes remain relevant across a wide range of application environments spanning everything from containers to blockchain databases. Most recently, much of those efforts have been confined to the various flavors of Linux that IBM makes available on its mainframes. Increasingly, however, IBM is signaling it expects customers will deploy emerging classes of workloads on z/OS as well. In fact, thanks to support for Kubernetes, it should become much simpler to move workloads between mainframes and other platforms.
It’s too early to say whether that capability will increase the number of mainframes being employed in the enterprise. But it is safe to say that whenever there is a mainframe present, the variety of workloads running on that platform is likely to be incredibly diverse.
