Home > Storage > PowerScale (Isilon) > Product Documentation > Storage (general) > PowerScale OneFS SmartFlash > L3 cache and performance
A scale-out storage system must deliver the performance required for various workflows, whether they are sequential, concurrent, or random. Different workloads typically exist across applications and, often, within individual applications. With OneFS and L3 cache, throughput and IOPS scale linearly with the number of nodes present in a single system. Due to balanced data distribution, automatic rebalancing, and distributed processing, OneFS can use additional CPUs, network ports, and memory as the system grows. This capability also allows the caching subsystem to scale in relation to cluster size.
One of the goals of L3 cache is to immediately provide solid benefits for a wide variety of workloads, including:
This goal is in line with the OneFS ethos of simplicity and ease of management. Although the benefit of L3 caching is highly workflow-dependent, the following general rules can be assumed:
Note: In archive-class nodes, L3 cache is enabled by default and cannot be disabled. In these platforms, L3 cache runs in a metadata-only mode. By only storing metadata blocks, L3 cache optimizes the performance of operations such as system protection and maintenance jobs, in addition to metadata-intensive workloads.