| Mapower MAP-TB32 Dual SATA RAID Enclosure
|
Testing (continued):
The two images below show the USB 2.0 performance for the SAFE-33 mode. The image on the left is benchmark data for the SPAN volume and the image on the right is benchmarking data for the RAID-1 volume. These numbers are strikingly similar to the previous benchmarks, but with marginally higher CPU usage. The one notable difference is access times for the RAID-1 partition, which is significantly lower than access times for the other modes. I suspect this is because the volume is smaller than the others and its partitions are located near the beginning of the hard drives.
 
The Mapower's USB 2.0 performance is average at best, I would expect the RAID-0 configuration to at least equal the Kingwin enclosures numbers.
The following benchmarks were performed using the TB32CS's eSATA connection and as you will see, this is the preferred method of connection.
The first two images show eSATA benchmark data for Mapower's TB32CS in SPAN (below left) and RAID-0 (below right).
 
The next two images show eSATA benchmark data for Mapower's TB32CS in RAID-1 (below left) and the Kingwin-EL-35EU-SBL (below right).
 
Note the large performance boost the RAID-0 configuration is able to achieve. Data throughput is higher across the board, but again this boost is achieved at the cost of higher CPU usage. Although, I should point out that it is only a marginal increase in usage. Performance data for the SPAN disks and the RAID-1 disks appear to be similar, but a look at the graphs shows that they achieved their results in different ways. eSATA performance is good, particularly in the RAID-0 configuration. All modes compare favorably with the Kingwin enclosure and the average transfer rate for the RAID-0 configuration is 25MB/sec higher.
The two images below show eSATA performance for the SAFE-33 configuration. The image on the left is benchmark data for the SPAN partition and the image on the right depicts the RAID-1 benchmark data. Again, these numbers are similar to the SPAN and RAID-1 benchmarks, but access times are lower and CPU usage is higher. The smaller partition sizes are most likely responsible for the lower access times and the higher CPU usage can be attributed to the combination of two different multi-disk modes that the enclosures controller is being required to manage. The only other item of note here is that the smaller RAID-1 drive in the SAFE-33 configuration doesn't suffer the transfer rate decline we see in the larger RAID-1 only hard drive.
 
The Mapower TB32CS performed slightly below the Kingwin enclosure in all but the RAID-0 configuration, but the differences are small and would be undetectable in general use. The RAID-0 mode significantly outperformed the Kingwin enclosure, which was expected.
|
|
|
|
