Improved system performance

Using GPFS™ file systems can improve system performance in the following ways.

Achieving high throughput to a single, large file requires striping data across multiple disks and multiple disk controllers. Rather than relying on striping in a separate volume manager layer, GPFS implements striping in the file system. Managing its own striping affords GPFS the control it needs to achieve fault tolerance and to balance load across adapters, storage controllers, and disks. Large files in GPFS are divided into equal sized blocks, and consecutive blocks are placed on different disks in a round-robin fashion.

To exploit disk parallelism when reading a large file from a single-threaded application, whenever it can recognize a pattern, GPFS intelligently prefetches data into its buffer pool, issuing I/O requests in parallel to as many disks as necessary to achieve the peak bandwidth of the underlying storage hardware infrastructure. GPFS recognizes multiple I/O patterns including sequential, reverse sequential, and various forms of strided access patterns.

GPFS I/O performance may be monitored through the mmpmon command. For more information, see mmpmon command in IBM Spectrum Scale: Administration and Programming Reference.