In this post I publish my results of some performance testing of My Cloud after using it for a week.
I did these testing because I was initially not happy with the performance to write data to My Cloud, doing the testing I found some best practices for me, maybe it is of any use for someone else.
In short, I am very satisfied with my purchase, it fits my requirements for a shared filesystem with safepoint reliability in my local network.
Best practice tuning
I followed the best practices from this forum to tune My Cloud, like stopping the processes S86wdphotodbmergerd and S85wdmcserverd at boot time with crontab and changing the rescan interval of the media server (twonky) to 0 to stop te scanning every time a file moves.
Testing results
My WD My Cloud is connected with a 1 Gb switch.
I did the testing from a Windows PC by writing (copy) files (videos and images) to the My Cloud.
I found that it matter with what tool you copy the data to the MyCloud. I achieved the best results with FastCopy ver 2.11, Windows Explorer en Beyond Compare showed to be a bottleneck.
My first test is with a Windows netbook with a 100 mbps ethernet connection.
The test showed that the 100 mbps network is the bottleneck, the full capacity of the network was used.
It does not matter if I copy from the internal harddisk or from a USB 2.0 disk, max possible throughput was 11 MB/sec.
Max throughput on the wire is 12,5 MB/sec, with 20% protocol overhead this is this very good.
My second test is with a Windows PC with a 1Gbps ethernet connection.
The results of copying from the harddisk to My Cloud, I achieved a sustainable write throughput of 20 MB/sec.
I checked these results with CrystalDiskmark v 5.0.2 and I saw simular write performance results in other published benchmark results.
But I’m not sure that the My Cloud is the bottleneck here. SysInternals Proces Explorer showed a high utilization of one cpu core (80-95%). I will see if I can repeat the tests later with a PC with faster cores.
I could not achieve these results by copying from a USB 2.0 disk, here it stopped at 17 MB/sec.
The USB disk was not the bottleneck, I could copy form the USB disk to the harddisk with 32 MB/sec, this is normal for a USB 2.0 disk. Copying in parallel with two FastCopy sessions did not help: I got 8.5 MB/sec per session.
Proces Explorer of SysInternals showed that the bottleneck was a fully utilized cpu core of the Windows PC.
So maybe I can achieve better results with a faster PC, I will test this later.
I hope these test results are of any use for somebody else.