Abstract
This paper describes the development of a Real-Time monitoring framework for parallel processes. There are various parallelism
techniques in parallel execution of processes such as Bit-Level parallelism, Instruction-Level parallelism, task parallelism. Besides these
techniques, different hardware architecture of machines for processing like pipelining of processes, superscalar execution, data dependency,
resource dependency are there. By using any of these parallelism techniques or hardware architecture, Real-Time parallel processing
applications can be developed. Our objective is to develop a data- driven framework for these Real-Time parallel processing applications.
The framework monitors entire lifecycle of processes executing in that system and gather granular level data to conclude entire system
information. This information can be used to reduce the overall execution time, chances of deadlocks, production cost along with
improvement in efficiency, consistency, resource utilization. Soft sensors record the entire details of process during its execution like
execution time, waiting time, delay etc. throughout life of processes.