Methodology of profiling network operations in software for distributed information systems

Abstract

Network operations take up large amounts of resources in distributed systems and in this case, existing profilers can correctly construct the execution tree of the program code and correctly identify the runtime, so you need other methods that will allow finding bottlenecks in the. The article discusses various approaches on the time efficiency assessment software, and based on a new method for profiling distributed systems, also an example of its use on a system built with agent technology-relational mapping. As a demonstration of the proposed method there’s presented an example of the system, built using the technology of agent-relational mapping that performs network operations in distributed transactions. Thus showing that for the ease of using this technology we have to pay with lots of operations to monitor network transactions.