Advancement Through Research
Data is not just getting bigger but it is also getting more connected. Graphs are natural and flexible structures that model the complex relations owing to their efficient and mature algorithmic support. Need to model and analyze these graph structures is increasing and it is exciting to implement graph-based methods over real life large-scale models. There are many challenges that still need to be addressed which include handling the massive input data with complexities of volume, velocity and variety. Various graph-processing systems have been put forth for analysis and processing of large scale graphs. In this paper performance, scalability, usability and ease of implementation for Page Rank, Connected Components and Triangle Counting workloads is studied for current age, widely accepted distributed graph processing systems Giraph and GraphX
Graph processing, vertex-centric, Giraph, GraphX, Page Rank, Triangle counting, Connected components
