ABSTRACT
The distributed system is a system where multiple processors can communicate with each other based on their connectivity for a common functional goal. While working on real-time distributed systems, we play with a limited set of connectivity, usually, they are regular highly connected or mesh networks. Other approaches for connectivity have been least explored, particularly while considering the network connectivity as asymmetric or heterogeneous. The approach of connection between processors/nodes of the system/network may play a vital role in the overall execution of assigned tasks. The focus of this article remains toward throwing light on how the connectivity of processors of distributed systems (dealing with real-time tasks, that is, tasks with a deadline (soft/hard)) may influence their overall productivities. This is understandable that the performance of an entire system depends on the timely execution of individual unit real-time tasks. With due consideration to space/time complexity, every network may have its functional carrying capacity. Network as a system uses many algorithms for manning or administering the task execution in the process, they may differ with different network architecture. Taking forward the former statement, we probe whether there is any difference caused due to differing connectedness. Unlike usual regular architectures described in existing systems, three different types of nature-inspired connectivity have been considered (for real-time distributed systems): regular connectivity, random connections, and scale-free structure. How mentioned structures can influence the timely execution of tasks will be discussed here. With the help of a mathematical model, will deliberate the creation of mentioned structure and execution of real-time tasks on it. The considerable parameters to decide the applicable connectivity among processors are migration time, available processor searching time with tasks success and failure ratio. Moreover, earliest deadline first and rate monotonic scheduling algorithms were standard to be used for the scheduling purpose of real-time tasks in this paper.
