Benchmark test. Helical Scan vs. Linear Recording tape drives and their susceptibility to dusty environments.
By:Runar Angelsen,
Product Manager -Media
Tandberg Data ASA, Oslo, Norway
Who is your tape drives worst enemy?
As the importance of the vital business data that is contained in today's computer systems has become ever so clear to the millions of professionals relying on the availability of their data, so has the importance of proper routines for protecting said data. Today the largely most common way of protecting mission critical is to take backup copies of the data onto magnetic tape.
As a result of this steady increase in use of tape backup systems, two major trends is apparent:
- Just like every other part of the computer business the need for speed, capacity and reliability for tape backup devices never stops to grow
- Tape drives are now required to operate in the same environments as other types of computer equipment, i.e. from the dedicated clean and well ventilated computer rooms of data centers to the work floor of garages, pizza bakeries and quarrying industry.
What sets the tape backup devices apart from other components in the computer systems is the fact that these devices must operate without the protection of the hermetic seal that protects the computers hard disk, microprocessor or other electronic components.
Hence the tape backup system must operate and perform to specifications while being subjected to intrusion of dust, fumes and other types of airborne contamination, your tape drives worst enemy.
Different tape backup technologies, how to choose
Today's tape backup devices ("drives") are being built around one of two technology platforms, linear recording or helical scan technology.
Helical Scan
The helical scan tape backup drives has their origin in video camcorders and DAT (Digital Audio Tape) recorders. Historically, the helical scan drives has provided higher transfer rate than comparable linear tape drives, at the cost of a much more complex and delicate design and lower reliability.
Linear Recording
As opposed to helical scan technology, the tape drives based on linear recording technology have a much less complex construction, Notably is the much less complex tape path and simplified mechanical design that provides robustness and reliability.
The important difference
Beside the obvious difference in mechanical complexity between Linear recording and Helical Scan tape drives, there is one more important factor in regards to susceptibility to dust, the interface between the recording head in the drive and the magnetic media inside the cartridge.
A linear recording tape will pass over an edge that will actually remove the dust before reaching the head. A helical scan tape will gradually be wound around a rotating drum causing dust to be dragged in between the tape and the head. This will cause the dust to damage both the tape and the head and reduce reliability and performance.
It should be noted that some helical scan tape drives (Sony AIT-1 and AIT-2) have a built-in automatic head-cleaning device. This device will however be activated after the drive detects an operational problem, and try to correct the problem rather than prevent it.
Figure 1. Difference in the head/tape interface between linear recording and helical scan
Third party tape benchmark dust test.
In the fall of 1999 Tandberg Data requested a third party test house to perform a benchmark test of tape backup drives to assess their susceptibility to dust contamination. NTS•XXCAL in Los Angeles CA, USA was chosen as an independent third party test house to perform the testing.
Purpose
The main purpose of the benchmark was to compare helical scan and linear recording tape drives with respect to the drives and media's ability to withstand dust contamination. Based on the recognition of dust intrusion as a main contributor to potential problems with tape drives, the difference in tape path complexity and mechanical robustness, such a test was expected to prove the more inherent robustness of the linear recording technology.
The test was set up using standard PC hardware and software similar to that found in normal entry level servers. The severity level (dust concentration and dust composition) in the test was chosen to simulate a normal office environment. All these steps were taken to simulate as closely as possible a real life situation.
Models tested
The test was performed using Tandberg Data's SLR (Scalable Linear Recording) tape drives, model SLR24 and SLR50, along with helical scan technology based DDS drives from Hewlett Packard and Sony
Media for the SLR drives was supplied from Imation, and for the DDS drives from FujiFilm.
Note: The Sony DDS4 drive is an external enclosure model, i.e. the drive has extra protection from dust by being built into an enclosure with its own power supply.
Technology / Manufacturer / Product type / Model / Capacity* / Media Manufacturer / Tape ModelLinear Recording / Tandberg Data / SLR / SLR24 / 12 GB / Imation / SLR24 12/24 GB
Tandberg Data / SLR / SLR50 / 25 GB / Imation / SLR50 25/50GB
Helical Scan / Hewlett Packard / DDS3 / Surestore DAT 24i / 12 GB / FujiFilm / 125M 4mm 12GB/24 GB
Sony / DDS4 / SDT-S10000/ME / 20 GB / FujiFilm / 150M 4mm 20GB/40GB
* Native capacity, non-compressed
Table 1. Tape drives and media used in test
Definition of severities
Basically, dust tests consists of placing test objects inside a chamber and operate it while exposing it to some degree of dust. Various test standards call out for either the dust to be blown into the chamber and circulated by airflow, or to be released from the top of the chamber and letting it settle freely onto the unit under test.
For this particular test an international standard ([1] IEC standard 60068-2-68, method Lb) was chosen. This test is performed by letting free falling dust settle on the test objects without added (forced) airflow. This simulates the effect of dust accumulating on equipment located in an office or normal indoor work environment. By adjusting the amount of dust released into the chamber per 24 hours, a test time of few days can simulate the effect of several years of exposure to dust sedimentation.
10 years of use
The amount of dust used in this test was set to 6 g/m2 per 24 hours, and the test duration to 144 hours (6 days). This is equivalent to 10 years exposure to dust in a normal office environment. See Figure 2above for the complete calculation of the acceleration factor.
Test set-up
To make a valid comparison between the various tape drives, NTS•XXCAL Test Engineers utilized the test bed as diagrammed in Figure 3below. Each server platform was configured with it’s own unique tape drive (i.e. one server was configured with Tandberg Data ASA SLR24 Tape Drive, the second server was configured with a HP DDS-3 Tape Drive, etc.). All tape drives were tested in their “shipped” configurations. In addition, each tape drive was tested under the same environment (e.g. same mounting location, etc.). Tests were performed on an isolated LAN to rule out the possibility of additional network traffic affecting the testing results.
Prior to the start of testing, NTS•XXCAL contacted each of the respective competitive vendors in turn, informing them that NTS•XXCAL is performing a benchmark/performance with their product while giving them a chance to submit the latest available released software and firmware revisions applicable to the benchmark/performance.
Test Procedure
Note:Prior to testing, NTS•XXCAL Test Engineers developed a test script using Rational Visual Test v 4.0r to perform backup, appends and restores of EIA-TIA files.
General test procedures:
Setup Server Platforms: NTS•XXCAL Test Engineers setup four (4) server platforms with Microsoft Windows NT Server v4.0 for testing. For details on server configuration, see Table 3below. Each server platform was configured with one of each of the tape drives under test (i.e. server #1 was configured with a SLR24 Tape Drive, server #2 was configured with SLR50 Tape Drive, server #3 was configured with a Sony DDS4, etc.). NTS•XXCAL Test Engineers then installed appropriate drivers for use under the Microsoft Windows NT operating system.
Installation: NTS•XXCAL Test Engineers then installed Seagate Backup Exec for NT v7.2 onto each of the server platforms, and verified that the platforms were still functioning properly prior to proceeding.
Benchmark: The tape drives were operated in a continuous backup/restore mode and the required dust concentration began. The backup/restore functions were restarted every 1.5 to 2.5 hours (after one full backup or restore) depending on the drive speed, for the standard 8-hour day. These conditions were maintained for six (6) days.
- NTS•XXCAL backed up a tape cartridge fully using EIA-TIA files (See note below)
(Note: During backup, NTS•XXCAL would backup the tape cartridge fully if the tape cartridge is equal to or less than the server hard drive capacity; otherwise, NTS•XXCAL performed append backups until the tape cartridge was full.)
- Once the tape cartridge was full, NTS•XXCAL attempted to restore the tape cartridge onto the hard drive.
- NTS•XXCAL erased the tape and repeated from step a until the test either failed or 144 hour of testing has been performed
Figure 3. Setup of drives and computers for dust test.
Platform Name: / Dell Optiplex GX1Processor: / Intel Pentium II/400 Mhz
System Memory: / 192 MB
Operating System: / Microsoft Windows NT 4 Server with Service Pack 5
SCSI Adapter: / Adaptec AHA-2940UW Ultra Wide
Hard Disks: / Seagate ST19171N SCSI hard drive (2 unit),
Maxtor 90640D4 IDE hard drive (1 unit)
Backup Software: / Seagate Backup Exec for Window NT v7.2
Table 3. Test platform hardware and software
Test result summary
Linear Recording drives
Tandberg Data SLR24
- 144 hours: Test completed. No functional problems experienced during test
Tandberg Data SLR50
- 144 hours: Test completed. No functional problems experienced during test
Helical Scan drives
DDS3, HP Surestore 24i
- 8 hours 14 minutes: Backup operation failed. Drive cleaned and restarted
- 84 hours: Restore operation failed. Drive cleaned and restarted with new tape
- 144 hours: Test completed. Two failure incidents during test.
DDS4, Sony SDT-10000ME, drive no. 1
- 21 hours 40 minutes: Backup operation failed. Drive cleaned but is unable to continue test. Drive replaced
DDS4, Sony SDT-10000ME, drive no. 2
- 25 hours 11 minutes: Restore operation failed. Drive cleaned and restarted
- 34 hours 11 minutes: Restore operation failed. Drive cleaned but is unable to continue test. Test terminated
Figure 4. Dust test, result summary.
Summary and Conclusion
After being exposed to dust sedimentation equivalent to 10 years of operation in an office environment, Tandberg Data SLR tape drives operates flawlessly throughout the test, while helical scan based DDS drives experiences severe operational problems, ranging from need for head cleaning to unrecoverable hardware failure.
The test at NTS•XXCAL [3] has shown that when it comes to a tape drives ability to withstand the adverse effect from dust and particle debris, its most common enemy, the tape drives based on Linear Recording shows much better results than drives based on Helical Scan technology. The reason for this is one or several of these factors
- Linear Recording tape drives has a much less complex mechanical design than their helical scan counterparts. Basically, this means that there is fewer parts that can be affected by dust, hence fewer parts that can fail.
- The Linear Recording tape cartridges contain the magnetic tape itself inside the cartridge at all times. Only a small door in the cartridge shell allows the magnetic recording head to make contact with the magnetic media. This results in less exposure of the magnetic tape to the dusty environment, which again reduces the risk of media failure due to dust.
- Helical Scan media on the other hand is pulled out of the cartridge shell, lead through a complicated system of guides and capstan rollers before being wrapped around a rotating head. This means that a large portion of the magnetic tape is exposed to the dusty environment as soon as it is loaded into the drive causing the problem described in Figure 1above.
In short: A fact of life is that dust appears everywhere in our everyday life and we expect our tape backup drives to operate in that same environment without any problems. This test has shown that when it comes to handling the adverse effects of dust in the environment, Linear Recording tape drives as the Tandberg Data SR24 and SLR50 has far better capabilities than comparable Helical Scan tape drives.
A word of caution:
As this test has shown, linear recording SLR tape drives are able to operate properly while being subjected to an environment that creates operational failure in helical scan drives.
Even though the Tandberg Data tape drives were able to operate in an very dusty environment without any problems, this should not lead to the conclusion that normal preventive maintenance of a tape drive can be ignored. All tape drives should have their read/write heads cleaned periodically according to product specification, to prevent unnecessary performance loss.
References
[1]IEC 60068-2-68 (1994-08)
Environmental testing - Part 2: Tests - Test L: Dust and sand
Specifies test methods to determine the effects of dust and sand suspended in air, on electrotechnical products.
[2]IEC 60721-3-3 (1994-12)
Classification of environmental conditions - Part 3: Classification of groups of environmental parameters and their severities - Section 3: Stationary use at weatherprotected locations
Classifies groups of environmental parameters and their severities to which products are subjected when mounted for stationary use at weatherprotected locations.
[3]NTS•XXCAL
Test Report No. 9905003-1
Dust Test Performed on Tape Drives
Page 1 of 7