Plasmon UDO30D-SE
3. Performance
We'll start by looking at IOMeter's output for the UDO Desktop Drive with the file server access pattern. We allowed the test to run for a period of 5 minutes. We also ran the same test on a Western Digital 80GB EIDE hard disk and a Fujitsu MHV2060AT 60Gb 2.5" drive, and have provided the output from both as a comaprison. While the two hard drives have very little in common with the UDO drive other than providing storage, the comparison does provide a way to judge the UDO Drive's performance and as we will see, there are some surprises.
Operation | Plasmon UDO Drive | Fujitsu 60 GB mini disk | WD800JB 80GB HDD |
IOps Read IOps Write IOps |
17.19 13.72 3.47 |
87.39 70.05 17.34 |
155.13 123.92 31.20 |
MBps Read MBps Write MBps |
0.1831 0.14724 0.03586 |
0.951854 0.765295 0.186559 |
1.676123 1.331478 0.344645 |
Transactions per Second | 17.19 | 87.39 | 155.13 |
Times are in milliseconds | |||
Avg. Response Time Avg. Read Response Time Avg. Write Response Time Avg. Transaction Time |
58.15 62.28 41.82 58.15 |
45.75 48.23 35.74 45.75 |
25.78 26.75 21.90 25.78 |
Max. Response Time Max. Read Response Time Max. Write Response Time Max. Transaction Time |
15,665.09 1,000.07 15,665.09 15,665.09 |
313.38 313.38 244.58 313.37 |
105.56 105.56 84.36 105.56 |
Errors Read Errors Write Errors |
0 0 0 |
0 0 0 |
0 0 0 |
Bytes Read Bytes Written |
46,401,024 11,300,864 |
241,184,256 58,794,496 |
419,644,416 108,622,336 |
Read I/Os Write I/Os |
4,124 1,043 |
21,053 5,212 |
37,249 9,379 |
The Western Digital Hard Disk wins hands down, as we would expect it to. However, the UDO drive does clock in some impressive figures for an optical drive. Its Avg. response times are not that far behind those of either hard disk.
The following table however, should prove more interesting. It shows the UDO Drive's output for 3 separate Iometer configurations, set to 100% reading of 512 byte, 16k byte and 32k byte file sizes. Setting Iometer to 100% reads, more closely simulates the type of work the drive will be expected to carry out. Again, the tests were run for 5 minutes.
Operation | 512 bytes | 16k bytes | 32k bytes |
Read IOps | 264.10 | 135.42 | 67.94 |
Read MBps | 0.13 | 2.12 | 2.13 |
Transactions per Second | 264.10 | 135.42 | 67.94 |
Times are in milliseconds | |||
Avg. Read Response Time | 3.78 | 7.38 | 14.71 |
Max. Read Response Time | 150.48 | 186.21 | 116.67 |
Read Errors | 0 | 0 | 0 |
Bytes Read | 40,644,096 | 666,877,952 | 669,155,328 |
Read I/Os | 79,383 | 40,703 | 20,421 |
It is interesting to see the difference in performance between small files and larger files. At 16k and 32k, the throughput (Read MBps) does not differ by much with 2.12MB/s and 2.13MB/s respectively. For 512 bytes however, it drops to 130kB/s. A similar pattern appears with total amount of bytes read, where for 16k and 32k the figures are fairly close at 636Mb and 638Mb respectively, while for 512 bytes, the corresponding figure is only 39Mb.
The last test we put our UDO drive through was to copy a full DVD movie to the drive. We used DVD Decryptor to copy Gladiator from a Pioneer DVR 108 drive. The process was completed in 1:00:40 with an average read rate of 1,974 KB/s (1.4x) and maximum read rate of 4,167 KB/s (3.0x). As a comparison, the same process but with the copy being written to the Western Digital Hard Disk, had a duration of 00:24:40 with an average read rate of 4,857 KB/s (3.5x) and maximum read rate of 6,949 KB/s (5.0x).