Availability is driven by capacity and performance. Without consideration for matching the capacity of the storage size and configuration to the workload, availability becomes a guessing game. Consequently, there must be a sound basis for the existing storage capacities and configurations. However, taking reality into consideration, using the most overlooked work axiom - do it right the first time - the existence of a sound capacity plan for an initial configuration may not be the case. On the other hand, storage networking, SAN more than NAS, is a new solution and a learning curve always begins with mistakes. Therefore, this chapter begins with the assumption that mistakes have already happened , can easily happen, and in the best case, can be avoided.
The premise of availability for storage networks is the application of the five systems management disciplines in an iterative process. These will be discussed as they occur in a real-time continuum, but we'll also view them in a cyclic manner. The following illustration offers a peek into the activities that make up the cycle. As we can see, a capacity model and implementation may set the basis for the start of the cycle, but may not have much to do with the reality of daily data-center operations. Performance management is influenced by both configuration and change activities, and given its investigative characteristics, forms a logical beginning to availability.
The performance management activities are supported by external events that are tracked through problem and application management systems. All of these activities develop a synergistic relationship that cycles back to influence the capacity model. So you see, it is, and continues to be, an iterative process in providing available systems.
This chapter discusses these interrelationships as they are related to the SAN and NAS environments. By applying activities and concepts from systems and network management practices, the availability of storage networks characterizes their fundamental difference compared to existing direct-attached storage models. In addition, these discussions begin to demonstrate the immaturity and challenges still evident within this area of storage infrastructures . Focusing on both fundamentals of systems, and storage management, the scope of these discussions illuminate key issues and offer insight into providing availability in storage networking environments.