Test setup

The Thecus N2200 Turbo NAS is connected directly to a Marvell Yukon NIC on an Asus Z7S. The host is Windows 2008 standard server. Network settings have not been changed, thus results will be actual results any home user could expect to see. For storage two Samsung HD321J 320GB harddisks are used.

Performance

For each RAID level (JBOD, 0, 1) I ran ATTO and copied a lage file to and from the public network share.


ATTO shows that in JBOD maximum speed is not reached with large amounts of data. Storing smaller files results in nearly double the performance compared to sustained speed with a large file. Considering a SoC is used to power the N2200, performance is not bad at all.


RAID 0 is usually used for high performance storage instead of redundancy. Performance is slightly lower than JBOD, probably due to the RAID calculations.



RAID 1 is the slowest solution from the three, however RAID 1 does offer redundancy. The performance hit is not that large, though we can clearly see its effects in our results.

Seeing how RAID 0 performance is (slightly) worse than JBOD and is less reliable by definition I would not recommend using RAID 0 at all. JBOD and RAID 1 are the only viable options in my opinion. The performance hit from enabling RAID is not very large though still measurable, I personally would not consider it a factor in picking the RAID level. In the end the choice will be between half the storage and data redundancy. This comes to personal preference and the importance of the data stored.

Power consumption

Power consumption is measured at the power outlet using a simple watt meter.


When the device is off 4 Watts are consumed.


When the N2200 is in standby mode and the disks are spinned down power consumed is at 16 Watt.


Once you access the device and start giving it work power consumption doubles to 32 Watts. Of course this can be slightly different when different disks are used.