Home > Storage > PowerScale (Isilon) > Product Documentation > Storage (general) > PowerScale OneFS SmartFlash > Overview
Enterprises across the globe are witnessing an explosion in the growth of big data today. These businesses continue to face the management challenges of this avalanche of large unstructured data. The ability to deliver greater performance from their storage systems while maintaining data consistency and availability for wide range of workflows also increases.
The caching of data and metadata is a common practice used to deliver high levels of performance and low latency in storage systems. Caching typically means to keep the most frequently accessed data in memory. Doing so provides faster access by intelligently predicting how content will be accessed and the parts of a dataset that will be required.
In traditional scale-up NAS architectures, the cache on a filer head processor is available only to the volumes stored on that device. Large deployments of these systems can consist of hundreds of volumes. As such, the total of cache on these systems might be huge, but its effectiveness is limited due to its partitioned nature. In contrast, the OneFS caching architecture ensures that all storage nodes in a PowerScale cluster share each other’s cache. It also ensures efficient use of the large aggregate amount of memory (RAM) in a cluster, which can contain a file system larger than 68 PB. Because each node contains RAM, OneFS read and write caches scale linearly as the cluster grows in size.