Home > Storage > PowerScale (Isilon) > Product Documentation > Storage (general) > PowerScale All-Flash F210, F710 & F910 > OneFS 9.7
From a software perspective, PowerScale OneFS 9.7 introduces a significant leap in performance. OneFS 9.7 updates the protocol stack, locking, and direct-write.
Implementing a round-robin distribution strategy across thread groups has significantly reduced thread lock contention. Moreover, thread group numbers are optimized based on core count, which leverages higher core counts to enhance parallelism. This leads to an increase in OneFS's overall efficiency and performance.
Contention on turnstile locks has been reduced by increasing the value of Read-Write (RW) Lock retries. This adjustment permits locks to spin adaptively for an extended duration. It is particularly noteworthy that contention on turnstile locks for Non-Uniform Memory Access (NUMA) nodes incurs a high cost, making these improvements even more valuable in optimizing system performance.
In the context of NVMe storage nodes, writing operations are strategically executed around the journal for newly allocated blocks. This method circumvents potential bottlenecks associated with journal flushers, thereby enhancing the storage system's efficiency and reliability. Such an approach is crucial for maintaining high performance and preventing data processing and access delays.
When combined, all of the software updates provide tremendous performance improvements across workloads. The software improvements are further expanded with the hardware upgrades.