Methods for large image distributed processing and storage

This paper presents a novel approach to develop distributed data-driven storage systems. The core concept behind this approach is the implementation of LH* schema that allows highly available distributed data storage to be created. We propose the novel method of image data decomposition and distribution among storage nodes to enhance main characteristics of schema. This method is close to image pyramid structure (also known as MIP maps) that allows the entire large image to be tiled, cropped and scaled but is less redundant and allows the image data in system become manageable. The storage system provides multiple access to digital repositories of imagery data at required scale and accuracy. Image is decomposed into data blocks of several levels. The image with the required scale and resolution is reconstructed from the corresponding set of downloaded blocks on client's side. The developed method allows adopting the compressing, redundancy and security algorithms usage and algorithm redundancy on the block level. Thereby we can flexibly manage the redundancy, security permissions and resource consumption.