Graph Processing System for Network Science

Abstract

Big graphs processing is a complex task that requires computational power as well as special purpose software. This happens due to the irregular structure of the graphs because it is not known in advance the way how the graph will be stored in memory. Also, since the size of graphs simulating real-world objects can achieve millions of nodes and edges (and even more), its efficient processing can be complicated by long search operations for navigating in the graph. Another obstacle to efficient development of graph algorithms is the non-deterministic structure of many such algorithms. With all of the above, this paper is devoted to a high-level description of a system that allows to develop graph algorithms efficiently and increase productivity of software developer. This system can also be used to build efficient data structures for storing graphs, since it allows to separate the development of algorithms for processing graphs from the issues of their storage on a computing system. The developed system is intended for use in research areas related to the processing of large graphs with complex algorithms, for example, in network science.