Efficient Structured Data Access in Parallel File Systems

Parallel scientific applications store and retrieve very large, structured datasets. Directly supporting these structured accesses is an important step in provid-ing high-performance I/O solutions for these applica-tions. High-level interfaces such as HDF5 and Parallel netCDF provide convenient APIs for accessing struc-tured datasets, and the MPI-IO interface also supports ef-ficient access to structured data. However, parallel file systems do not traditionally support such access. In this work we present an implementation of structured data access support in the context of the Parallel Virtual File System (PVFS). We call this support “datatype I/O” because of its similarity to MPI datatypes. This support is built by using a reusable datatype-processing component from the MPICH2 MPI implementation. We describe how this component is leveraged to efficiently process structured data representations resulting from MPI-IO operations. We quantitatively assess the solution using three test applica-tions. We also point to further optimizations in the process-ing path that could be leveraged for even more efficient op-eration. 1