SuperDLT 320 outperforms first-generation LTO drives in benchmark tests.
By Jack Fegreus
The debate over tape cartridge formats is being driven by the need for increased density (capable of delivering library capacities anywhere from a few to tens of terabytes) and increased throughput.
In the midrange market, the two dominant tape formats are DLT and LTO. This review compares the performance of SuperDLT (SDLT) 320 drives to LTO-1 Ultrium tape drives. Second-generation LTO (LTO-2) drives were not available at the time of this review.
SDLT 320 technology
Quantum leveraged the characteristics of serpentine recording in a conservative manner to provide a dramatic boost in single-reel cartridge capacity. SDLT 320 introduced changes in two key design areas.
Compared to the previous generation of SDLT drives (the SDLT 220), an SDLT 320 drive is capable of writing data 45% denser on the tape. While an SDLT 220 drive lays out data in tracks at 133,000 bits per inch (bpi), an SDLT 320 drive packs data at 193,000bpi. In comparison, the data density of an LTO tape is 124,000bpi.
Figure 1: The OBLtape benchmark pegged the native (no compression) performance of the SDLT 220 drives at 10.7MBps and the SDLT 320 at 15.6MBps.
It is important to note that SDLT 320 drives use SDLT-I type tape. In other words, the SDLT 320 does not require any new media investment by current SDLT users. SDLT 320 drives are also backward read-compatible with DLT-type IV media, thereby bridging DLT and SDLT.
Along with greater density, Quantum wanted to provide tape heads that could deliver higher data transfer rates, but were less susceptible to environmental conditions than traditional heads. It also wanted to improve head life. To this end, Quantum introduced Magneto-Resistive Cluster (MRC) read/write heads. The tape drive also implements advanced Partial Response Maximum Likelihood (PRML) channel technology, which is used by many disk-drive manufacturers.
With PRML, the channel compares the measured signal from the tape with a known waveform in order to interpret the data. PRML attempts to interpret even small changes in the analog signal, whereas peak detection relies on fixed thresholds. As a result, a drive using PRML can correctly decode weaker signals and read/write data at a higher bit density. PRML increases the tape bit density by 45%, speeds up the tape movement from 116 inches per second (ips) to 122ips, and enables a sustained native throughput of 16MBps.
With the tape streaming by the heads at 12ips, Quantum also addressed technological hurdles to ensure continued low bit-error rates as areal bit density increases in future generations of SDLT. To this end, SDLT drives also introduced an optical servo technology called Laser Guided Magnetic Recording (LGMR).
SDLT marries optical laser servo technology with magnetic write/read technology. LGMR ensures even higher cartridge capacity by servoing from optical targets on the media's backside. As a result, 100% of the magnetic surface is dedicated to reading and recording data tracks. This compares to traditional magnetic tape designs, which reserve 10% to 20% of the recording surface for embedded servo track information.
Quantum's solution was to laser-etch optically readable servo tracks on a specially formulated back coating of the media and to use a three-beam hologram configuration for exact tracking. Furthermore, the laser servo tracks cannot be magnetically erased. This indelible servo information eliminates the need to magnetically pre-format tapes. It also makes bulk erasure of Super DLT cartridges possible. (Traditional media must have the servo tracks re-recorded after a bulk erasure.)
PRML and LGMR technologies ensure faster, higher-capacity, and more-reliable tape drives. SDLT 220 cartridges held 110GB (non-compressed) of data. With the SDLT 320 drives, those same SDLT I tape cartridges can hold 160GB (320MB, assuming 2:1 compression).
To test the performance of the SDLT 320 drive, OpenBench Labs calibrated the drive using its OBLtape v1.0 benchmark. We ran the tests on the same hardware running under SuSE Linux (with the 2.4.18-4 kernel) and Windows 2000 Server.
Our tape benchmark generates two very different types of data streams: purely random data and data that falls into a preset frequency (compression) pattern. The benchmark's patterned data stream was originally calibrated using Quantum DLT7000 tape drives, which implemented the Digital Liv Zempel (DLZ) compression algorithm.
That algorithm provided approximately a 2:1 compression ratio on normal data. OpenBench Labs devised a means of generating patterned data that consistently produced a compression ratio on the order of 1.9:1 to 2:1 on those DLT 7000 devices.
The OBLtape benchmark starts by allocating a large block of memory from which it then streams patterned or random data to the device. By streaming data directly from memory, the benchmark eliminates bus-bandwidth contention with other devices. In this way, the benchmark more accurately represents the data transfer rate of the tape device than the overall system data throughput. The data can be streamed in block sizes of 2n KB, where n ranges from 0 to 8. This simulates the differences in the way backup applications read data off a disk drive. In particular, high-end backup applications tend to use 64KB reads on Windows and often use 128KB reads on Linux. For that reason, we chose to use 64KB data blocks on tests run on Windows 2000 Server and 128KB data blocks on Linux tests.
We used a QLogic Ultra160 SCSI host bus adapter to connect to the tape drives. The testing was conducted on a dual-processor Hewlett-Packard Netserver 1000.
The results proved to be interesting on a number of levels. First, there was some expectation that there would be differences in the streaming performance of each tape drive between the two operating systems. For all practical purposes, there was none.
Using fresh media, with data compression turned off and using 128KB data packets on Linux, we measured native throughput of the SDLT 320 at 15.6MBps. Using 64KB data packets on Windows 2000 Server, the throughput was identical. Turning compression on and sending packets of patterned data brought throughput up to 32.1MBps on both Linux and Windows 2000 Server.
These performance results pegged the SDLT 320 as the highest-performance half-inch single-reel cartridge drive. (Editor's note: LTO-2 drives were not available at the time of these tests.) The Seagate Viper 200 had a modest edge in overall performance for the LTO drives. Nonetheless, it trailed the performance of the SDLT 320 in all categories. Normal uncompressed throughput was measured at 14.2MBps on the Seagate Viper. With compressible data, throughput on the Seagate drive held steady at around 28.7MBps.
In the next issue of InfoStor, OpenBench Labs will test an LTO-2 drive from Hewlett-Packard.
Jack Fegreus is lead analyst for the CCI Group, which conducts research studies on the functionality and performance of IT software and hardware. He can be contacted at JFegreus@CCIcommunications.com.
Half-inch tape cartridge drives
What we tested
Quantum SDLT 320
HP Ultrium LTO
How we tested
SuSE Linux v8.0
Windows 2000 Server
QLogic QLA12160 HBA
OpenBench Labs OBLtape v1.0 benchmark