Optimal real-time DB management

Abstract

A different processor paradigm for real-time command and control problems is presented, using the air traffic control problem as an example. Current ATC approaches use "dynamic" scheduling algorithms that seem to guarantee unpredictable performance. The current problems with ATC were pointed out, since 1963, by theoretical results in real-time scheduling. M. Klein et al, state, "an efficient real-time multiprocessor scheduling algorithm is not expected" [2]. J. Stankovic et al, write "complexity results show that most real-time multiprocessing scheduling is NP-hard. "[3]. Note that each author is discussing multiprocessing. The associative processor (AP), a poorly understood set processing Harvard architecture, offers a feasible, predictable and economical approach to real-time database management problems and avoids the problems that are inherent in present day multiprocessor algorithms. We further consider the rationale for this performance by presenting a new time complexity function that considers set processing in the real-time database environment. Following Garey and Johnson in their renowned book "Computers and Intractability: a Guide to the Theory of NP-completeness" [4], we develop the rationale for a time complexity function 0(1) for n operands. Because of the Harvard architecture, only a single instruction can exist at any time. When the summation of algorithm times is less than the allowed update time the schedule is optimal.