NOTE: This topic is part of the Uptime Information Hub.
Isilon scale-out network-attached storage (NAS) is capable of scaling to 144 nodes and 50 PB of unstructured data in a single, space-sharing namespace. Optimal management of this storage asset will result in:
- Stability: Accommodates workload “burstiness” without performance impact
- Continuity: Offers a business acceptable window of risk during maintenance activity
- Latency: Assures acceptable user response time
- Throughput: Meets expected throughput numbers
- Dependability: Allows jobs to finish in the allotted time
- Reliability: Causes no escalations that divert staff from other projects
To achieve these business goals, storage administrators must understand the principles of “performance” variation across ranges of capacity utilization.
Many customers seek guidance on balancing performance against full utilization of cluster capacity. The Isilon OneFS Cluster Optimization through Capacity Management white paper investigates the effects of capacity utilization on performance and cluster reserve capacity best practices.