Rapidly merging and sorting algorithm for isomorphic big data in distributed application system

Some big companies would deploy application sub-systems with same logical functions in different domains in order to expand business and maintain synchronization in business, such as deploying power marketing system 2.0 in the State Grid corporation. And those sub-systems in different domains will produce big data which are the company’s good data assets and can provide data support for the overall analysis of the company. So, this paper provides a rapidly merging and sorting algorithm for isomorphic big data in multiple domains to finish big data from different sub-systems to gather in one big data set to deeply analyze. This algorithm based on a 2-way merging method joins the dynamic merging strategy to reduce data waiting time; and then the algorithm is improved by melting the parallel merging approach to boost the merging speed. Further, in order to sort the merging data, the algorithm involves a custom inserting sort. Finally, this algorithm has been proven to rapidly merge and sort data to rise at least 50% compared with the simple merge sort method through experiments on the simulated marketing big data in the State Grid in China and can provide an important reference for rapid merge sort of distributed big data.