Test protocol
To measure the performances of these NAS, we used the following components:
– five Seagate 500 GB (ST3500320AS) hard drives installed in every NAS including the Buffalo in order to obtain comparable results (especially concerning power consumption as nothing guarantees that a Terastation is equipped with the same drives as another Terastation.) ;
– a Netgear GS-108 gigabit switch which manages Ethernet packets up to 9 KB (jumbo frames) ;
– an Enermax Jazz hard drive casing whose particularity is to offer a double USB 2.0/eSATA interface;
– a sound probe, thermometer and ammeter to measure physical parameters.
Rather than synthetic benches like Iometer or Iozone, we decided to look at the time it took for simple file copying. At least two reasons explain this choice. On the one hand, this gives us transfer speeds that are much more representative of the real use of these machines than theoretical measurements. On the other, it also allows testing, in a strictly identical manner, the different access protocols: essentially SMB/CIFS and FTP, but also iSCSI and external storage connected on the USB/eSATA ports.
Three groups of files were used in tests :
– large sized files (two files totaling 1.83 GB), two Divx movies;
– medium sized files (63 files totaling 1 GB), a mix of MP3 and Flac files ;
– small sized files (14,746 files totaling 659 MB, all inferior to 1 MB), a website!
Tests were carried out in reading/writing with each measurement taken three times. For CIFS/SMB, a simple batch Robocopy aided us in measuring file copy times. For FTP, Flash FXP logs gave us the transfer times.
File transfers were launched using the following configuration:
– Core 2 Duo E4500 ;
– 2 GB of DDR2 667 MHz memory ;
– Asus P5K-SE motherboard ;
– two Western Digital WD5000AAKS-0 hard drives in Raid 0 for reading and the copying of files (speed in writing measured at 75 MB/s) in addition to the system drive ;
- a Gigabit D-Link DGE-530T card (PCI, Marvell Yukon circuit).
All NAS were tested with the latest available firmware although there were two exceptions. For Synology, we used the 518 version (stable) for performance tests and the 571 (beta) to test functions. The latter adds significant innovations and we had to include it.
As for Qnap, we used the 20080122 version for most tests. This is the latest official version ; however, it functions very poorly for the FTP transfer of small files. After having brought this to Qnap’s attention, they provided us with the beta version of its firmware 20080219 which corrects the problem, but it is still very unstable in tests.
