tailieunhanh - THE FRACTAL STRUCTURE OF DATA REFERENCE- P10

THE FRACTAL STRUCTURE OF DATA REFERENCE- P10:For purposes of understanding its performance, a computer system is traditionally viewed as a processor coupled to one or more disk storage devices, and driven by externally generated requests (typically called transactions). Over the past several decades, very powerful techniques have become available to the performance analyst attempting to understand, at a high level, the operational behavior of such systems. | Hierarchical Reuse Model 3 1 Figure . TSO storage pools distribution of track interarrival times. Probability exceeded Probability exceeded Figure . TSO storage pools distribution of record interarrival times. 32 THE FRACTAL STRUCTURE OF DATA REFERENCE Figure . OS 390 system storage pools distribution of track interarrival times. Figure . OS 390 system storage pools distribution of record interarrival times. Hierarchical Reuse Model 33 single-reference residency time lengthens. Thus as we should expect these plots suggest an important role for processor file buffers in the production database storage pools. In Chapters 3 and 5 we shall sometimes adopt a mathematical model in which multiple workloads share the same cache or processor buffer area and each individual workload conforms to the hierarchical reuse model. This results in a series ofequations ofthe form one for each workload. In graphical terms it corresponds to fitting each workload s plot of interarrival statistics with a straight line. Collectively Figures and 1 through provide thejustification for adopting the mathematical model just described. The multiple workload hierarchical reuse model as just outlined in the previous paragraph is both sufficiently simple and sufficiently realistic to provide a practical framework for examining how to get the most out of a cache shared by multiple distinct .