Vector data processing by means of specifications in accordence with the MapReduce programming model

One of the approaches to accelerating data processing is the application of the distributed computing model MapReduce. In this model, input data are distributed among the computational nodes, processed, and then assembled into the resulting body of data. Existing software that implements MapReduce does not take into account the specificity of spatial data processing. In order to minimize the service execution time and to provide uniform use of available computing resources of the local cloud infrastructure of SB RAS Institute of System Dynamics and Control Theory within the framework of the MapReduce model, we have proposed the implementations of operations controlled by specifications for parallel execution of services. The article considers application of specifications for processing the results in a vector format on the example of a web service for object identification.