Execution Trace Streaming Based Real Time Collection of Dynamic Metrics Using PaaS

The large amount of data that needs to be processed, still remains one of the major hindrance toward widespread acceptance of dynamic metrics. This paper presents a live execution trace streaming based approach using platform as a service (PaaS) to overcome this hurdle. The proposed approach is the first of its type having the ability to provide the users with live dynamic coupling metric trends. Experiments were carried out on five sample applications with varying size ranging from 100 to 1250 classes. The performance of proposed approach remains to a large extent, independent of increase in application size, number of classes and methods used, number of messages exchanged, and execution time.