Autonomic computing features on the zSeries are the culmination of a 40-year heritage that began with the introduction of MVS, and the ability to run mixed workloads on a single server. The S/390 of the ’90s first introduced Workload Manager (WLM) and Parallel Sysplex clustering; these technologies are the basis for many of the autonomic computing enhancements found in zSeries today.
This article reviews how IBM has consolidated autonomic computing advances over the years on the zSeries and looks at future zSeries direction for autonomic capabilities.
Autonomic Computing Origins
The IBM zSeries has the most continuous history of autonomic computing of any IBM platform. In fact, many zSeries users may not even be aware of the various autonomic features incorporated in the machine since its inception in 2000.
For example, the IBM Server Group’s eLiza initiative continues to exist in the zSeries as a specific implementation of autonomic computing technologies. Today, however, eLiza is known as a set of technologies that will be common across eServer platforms and that can differentiate the eServer line from competing lines. This same internal IBM server group has had a particularly strong focus regarding the successful establishment in their product lines of what’s now classified as autonomic functions or constructs, and that also includes technologies such as heterogeneous workload management, a trademark of zSeries computing.
Now endowed with the powerful z/OS flagship operating system, the zSeries comes equipped with even more automated decisioning for resource balancing and allocation. It reduces the demand on critical IT personnel and the costs of system operation. In continuing to arm the zSeries with new autonomic capabilities, IBM clearly intends to eventually deliver totally hands-off performance management across all systems and storage in an enterprise.
Autonomic Computing Value Proposition
Mary Moore, IBM z/OS Marketing manager, comments, “Autonomic computing, which is the concept of a self-managing system, addresses two critical IT requirements: cost and complexity of operation. The purpose of autonomic computing is to drive functionality into products.”
Today’s corporate IT manages a vast array of systems—from customer relationship management (CRM) to Web ordering— as organizations look to consolidate servers. While IT works to consolidate diverse systems and departmentalized servers into more centralized architectures, workloads at the server, network, and network device levels all need to be rebalanced and managed in a total enterprise picture. There are several inherent challenges to this.
First, IT professionals lack empirical knowledge on best practices for resource usage and distribution on servers and modern networks. This contrasts starkly with the old mainframe days, which featured years of experience in system resource balancing and management do’s and don’ts. Much of this lack of best practice information is due to the informal and distributed means by which server-based computing has entered organizations. User departments have purchased many of these resources, usually without thinking about optimizing system resources and often without consulting IT.