On Information Theoretic Entropy of Real Time Distributed Systems

Abstract

Utilization factor is the most significant feature of Real Time and Distributed System In a DS if any task requires migration or duplication, first scheduler checks the processor utilization and then duplicate it towards the destination processor Likewise, in RTS the acceptance test of every scheduling algorithm is determined by using utilization factor and#119875;and#119903;and#119900;and#119888;and#119890;and#119904;and#119904;and#119900;and#119903;and#119880;and#119905;and#119894;and#119897;and#119894;and#119911;and#119886;and#119905;and#119894;and#119900;and#119899;1 Earliest Deadline First Rate Monotonic Scheduling Least slack time first Pfair are some renowned scheduling algorithms of RTS They usually schedule real time tasks on the basis of utilization imposed by task upon respective processor However, the unexpected arrival and execution of tasks create bafflement in the system This bafflement generates overloading which causes tasks to start missing the deadline This bafflement is nothing but uncertainty present in the system; if we quantify the amount of uncertainty, then this uncertain task overload can be regulated precisely and efficiently This amount of uncertainty can be quantified by using the proposed information theoretic concept of entropy in RTDS Hence, the primary focus of this thesis is arrival of entropy in RTDS as a new guidance parameter Along with the introduction of this novel parameter, we also analyze new task duplication and migration based scheduling algorithms for DS and RTDS