Hello fellow spiceheads.
I want your help to better access the impact of network equipment on the access speed of large file sets over the LAN.
Clients that rely on Solidworks for planning complex machine models access thousands or even tens of thousands parts in a shared library over the LAN, each part represented by a single file. Large projects can take a few minutes to open on a gigabit network (L2 managed switches) on a quite fast Sinology NAS.
I know both disk access times, type of IO (small x large files) and switch speed affect the overall fileset opening time. Therefore, I have downloaded ATTO disk benchmark which works on mapped network drives as well, and used its benchmark tests to have real world NAS performance results (attached). Please check the comment box for more info on the different test environments (both local and over NAT).
A preliminary test shows a linear speed bump in proportion to file size both on NAT and locally. A weird finding is the incredible speed of a RAID10 controller that has a 256MB buffer with backup battery. Actually, results are unbelievable... :)
Can you please add your own test results to this thread ? I am more interested on the effect of different types of switches, including cisco, 100Mb, etc. The more the merrier.
THANKS !