Computer systems depend largely on their keepers, whether they be embedded systems, the latest blade servers, or the largest super computer. Computer operators and systems programmers have largely given way to increased areas of specialization for administration, maintenance, and servicing . Although many mundane tasks have been automated, more complex configurations of distributed systems have increased the need for system-oriented personnel to monitor and keep the servers running. This is because, no matter how sophisticated they become, computer systems remain electronically controlled components , susceptible to both human and mechanical error.
The basis for these activities revolves around the accepted disciplines of managing systems. These are known as the five systems management disciplines and can be applied to any macro system in operation. These disciplines include the following:
Performance Management The activities associated with the monitoring, tuning, and analysis of a computer system's (or subsystem's) performance.
Capacity Management Analysis activities that drive the planning, design, and implementation of computer resources necessary to maintain established service levels.
Configuration Management Activities that surround the design and implementation of a computer system; this includes ongoing support of system enhancements and upgrades.
Change Management Driven by performance and configuration management activities, these practices monitor and supervise changes that occur to computer systems.
Problem Management Activities that track, analyze, identify, and resolve system problems.
Before you object to the absence of applications management, security management, network management, web management, and others, I should first explain how these topics relate to the subject. The basic disciplines of systems management, as discussed previously, can be applied to any logical definition of a system. Consequently, targeting business applications for management can, and should, have the same disciplines applied to it-in other words, performance, capacity, change, configuration, and problem. By any other name , they usually do. However, particular topics, such as security, add a different dimension to management practices given they are so integrated with the supporting infrastructure. Security, which will be discussed in Chapter 25, should be regarded as a supporting system, like backup and recovery, or database administration. Even here, foundation disciplines can, and should, be applied.
As discussed in Chapter 1, the difference between enterprise applications and programs and utilities that support the systems' infrastructures cannot be overlooked. They form synergistic relationships between the productive and the necessary. Applications drive the business and form the primary components for workloads. However, the performance of computer systems depends on the effective interaction between productive work and overhead. Although many systems programs and utilities are referred to as applications themselves , they often offer productivity enhancements to the systems management activities and therefore to the effectiveness of the IT staff. However, they do nothing for the bottom line. In analyzing basic systems management disciplines, its important to remember that they do not exist purely on their own. Without the bottom line application of computing for some productive processes, the activities that monitor, track, and analyze computer systems would be non-existent.
Although we often consider the management of business applications to be the pinnacle of systems management, there are many other scientific, academic, and government endeavors that require activities just as stringent as general business computing. Consequently, the five systems management disciplines are not aligned to any vertical market or work orientation.
Computing continues to be a general-purpose discipline where applied technology and operational practices migrate within the society of computer professionals who perform management operations. Practices that worked well in scientific areas have migrated into business over time, with both government and military agencies using unique database management, just to name two.
The point is that computer system management is an iterative process and continues to evolve through cross-pollination with different application bases, such as storage networking. Given that storage networking is a different storage model, however, existing management practices, though relevant, need to be molded to support a new infrastructure, which is what storage networking creates, a new application of storage.
In evaluating management of storage networking, we must apply the five disciplines. Figure 21-1 shows how these trans-technology practices are applied to the ever-increasing diversity within the data center further separating the systems into specialized practices. However, as Figure 21-1 also points out, storage networking draws heavily from data center practices and interfaces with each in order to be an effective infrastructure.
Planning computer systems to meet business continuity requirements has been one of the most tedious jobs within Information Technology. In the past 30 odd years of established data centers, business continuity responsibilities were ill- funded exercises in theoretical computer systems planning. However, as the world's socio-economic conditions began immersing computer systems within business operations, whether for the sake of competitive positioning or national defense, the discipline of disaster recovery became a more imperative science.
Computer storage has always been the linchpin of any recovery scenario. From legacy systems to existing blade level embedded servers, all are linked to the vital circulation of data. Without it, the computer infrastructures they create would be worthless. Historically, the storage component was considered part and parcel to the holistic system that supported the evolving set of automation we have grown to depend on. Only recently has this link been recast into an infrastructure of its own. Storage networking in its inherent value has caused many to rethink their data center continuity plans, change their processes, and modify and enhance their solutions according to this new storage paradigm.