The main purpose of the performance review is to provide our customers with an estimate of performance of FlexTk's built-in
DOD 5220.22-M compliant data wiping on different hardware configurations and data sets. In addition, we have compared
FlexTk v2.8.40 to other popular data wiping solutions such as
XL Delete v1.7.2, Wiperaser v7.0.1 and BCWipe v3.11.
As you can see on the average results graph, FlexTk is capable of wiping data at a speed of 6,380 KB/Sec on average,
BCWipe is on the second place with 3,973 KB/Sec, XL Delete is on the third place with 2,385 KB/Sec and Wiperaser is
the last with 1,145 KB/Sec only.
FlexTk's data wiping engine is optimized for multi-core/multi-CPU computers and advanced RAID storage systems and capable
of wiping huge amounts files very fast and effectively. In order to accurately measure the performance of all the selected
data wiping tools, all tests were performed on the following storage devices:
- WD Raptor 150GB
- 2x WD500GB RAID0
- NAS Server with WD500GB over Intel Gigabit Ethernet
- WD500GB External USB Disk
All tests were performed on a powerful PC machine equipped with a Q9450 Quad-Core CPU, 4 GB of system memory and running
the Windows XP SP3 32-Bit operating system. Data wiping over network was performed from the test machine to a powerful
NAS server interconnected through a pair of Intel's Gigabit Ethernet NICs.
In order to measure a statistically representative level of data wiping performance, we have prepared a diversified file
set containing 2,000 (300 MB) files including:
- 150 MB of medium-sized MS Word documents (50K-1M)
- 150 MB of large image files (3-5MB)
Each data wiping tool was executed on all hardware configurations using the 7-Pass, DOD 5220.22-M compliant data wiping
algorithm and the same file set. In order to completely eliminate OS file caching effects, the test machine
(and the NAS server during network tests) was rebooted after each benchmark. Individual benchmark results from all four
different hardware configurations were averaged and finally five graphs (average and four hardware setups) representing
all the performance results were prepared.
Testing data wiping performance may be a little bit tricky as it is not evident whether faster tools are performing all
the wiping operations required by the DOD 5220.22-M compliant data wiping algorithm, which should overwrite all the data
seven times using specially designated data patterns. There is always a potential that some software vendors may reduce
the number of required data wiping passes in order to make their data wiping tools faster.
In order to test this specific issue and to verify whether all the tested tools are performing all the required wiping
passes, we have decided to perform two additional tests. The first test is to check the network link utilization during
data wiping process over network and to see at what speeds tools are actually pushing data over the network.
As it is clear on the network utilization graph, the data wiping performance results are completely consistent with the
network link utilization for each specific tool. Fast tools are pushing data faster and slow tools are just slowly writing
data to the disk.
During the second test we have measured the total amount of data that has been written to the disk during the data wiping
process for each specific data wiping tool. As it was already mentioned, a DOD 5220.22-M compliant data wiping algorithm
should overwrite all the data 7-times and therefore we should expect about 2.2-2.4 GB of data to be written to the disk
during the data wiping process for our specific file set (300MB x 7 + overhead).
As you can see on the last graph all the tested data wiping tools write about the same amount of data to the disk and all
of them are well above the expected amount of 2.2-2.4 GB of data. The only conclusion we can get to is that all the tested
data wiping tools are actually performing all the requited wiping passes, but some of them are just doing the operation
faster than others.
* This performance review has been prepared for information purposes only and we are strongly advise
you to make your own performance evaluations using your specific hardware components and datasets.
