SHORE is a middleware framework designed to facilitate the implementation of plugins to support a range of storage and I/O interfaces on the backend side and a range of data format plugins on the frontend side. This allows a separation of I/O optimisation from application optimisation. SHORE has been desinged and implemented as part of a PhD thesis and is thus not production ready software, but more a proof of concept which in fact has been used and deployed for real-world science data reduction on very large scales.

SHORE is not an acronym, but can be loosely interpreted as Storage Hierarchy Optimization and Retrieval Environment. The idea of SHORE comes from the scene where waves hit a seashore and disappear in the sand grains. This is similar to a piece of data 'disappearing' in some storage. The concept of the shoreline is the interface between the ocean (data) and land (storage). A shoreline extending to distance represents a horizontally scalable data I/O interface.

When a wave hits the shore, water fills from top down. Similarly, in a data ingestion process, storage hierarchies with higher throughput are usually desired first. Then data is moved to slower hierarchies when faster ones are filled up. Sand grains represent plugins for a certain data format or type. The poles of a jetty going out into the sea are sensors picking up the dynamics and type (data type) of the water (data) before it hits the shoreline, informing the shoreline where to finally put the water (which plugins to use).

Waves (data ingestion) can also come in different intensities, high (intensive) ones and low (sparse) ones. In order to handle each type in the optimal way, there needs a data-aware plugin infrastructure, to make decisions dynamically based on characteristics of a data ingestion. In the opposite direction, a storage-aware infrastructure (jetty) is also necessary to provide the data object (wave) with storage hierarchy (land) information before the ingestion happens. Apart from the main architecture, there are also concepts for optional functionality, such as shells, which provide the in-situ data processing ability.