Scheduling distributed applications can be challenging in a multi-cloud environment due to the lack of knowledge about the application characteristics. In order to realize a versatile multi-cloud scheduling algorithm,knowledge about the application’s runtime behavior over various resources is needed. Besides,not all applications exhibit the same kind of resource consumption pattern at all stages. Thus,looking into the resource consumption pattern,extracting the knowledge and classifying the applications can help in better decision making in a multi-cloud environment. The aim of this project is to create a profiler component that monitors the resource consumption of applications and stores it in a profile database. This profiler monitors the whole workflow,so that any user constructed workflow’s execution can be estimated by the use of certain Machine Learning algorithms using which an Analytical Model can be created and the estimation is carried out using the profiled data. Using the profile database,the execution time of an workflow can be estimated based on the resource consumption data of the workflow. In this project,we profile certain scientific applications and monitor their resource consumption pattern on different hardware configurations. After profiling,the applications are classified as being memory-intensive,cpu-intensive etc and using the resource consumption pattern of the application,execution characteristics of a particular workflow can be predicted. This is done with the help of certain machine learning algorithms like k-NN classifier,k-NN regression and SIGAR API is used for extracting resource consumption data. © 2016,International Journal of Pharmacy and Technology. All rights reserved.