Next-gen tape battles: Exabyte's Mammoth-2

Posted on July 01, 2000


by Jack Fegreus

For the next generation of enterprise-class tape systems, the battle lines are being drawn among LTO, backed by HP, IBM, and Seagate; SuperDLT, backed by Quantum; and Mammoth-2, backed by Exabyte. While LTO and SuperDLT will not debut until later this year, Exabyte's Mammoth-2, which sports a native uncompressed thoughput rate of 12 MBps on writes, has already begun shipping. But that fact is the least significant difference among these high-end tape alternatives.

Mammoth-2 is the next generation of Exabyte's 8-mm helical-scan Mammoth tape-drive line. As such, it continues to underscore the fundamental difference between helical-scan and linear tape technologies as currently represented by Quantum's DLT and the soon-to-be released SuperDLT and Ultrium line of LTO. To drive throughput, the current generation of linear tape drives move tape rapidly past stationary heads at a speed of up to 160 inches per second (ips). In contrast, helical-scan drives rely on a slow-moving tape-1.8 ips-crossing a fast-spinning set of heads mounted on a drum or scanner. The net result is a relative head-to-tape speed for the Mammoth-2 of 547 ips.

Figure 1: The Exabyte Mammoth-2 drive proved to have a base performance four times greater than the new generation of helical-scan DAT drives. With hardware data compression, peak throughput was measured at 22.8 MBps.
Click here to enlarge image

This fundamental difference naturally results in very different operating characteristics. As the name implies, DLT and LTO write data in long parallel tracks that run the length of the tape. Filling a tape with data requires numerous passes as the tracks wind in a serpentine fashion. In the case of the DLT 7000, it takes 52 passes to fill the tape. Helical-scan technology, however, writes data in short angled tracks that run across the width of the tape. Because the axis of rotation is not orthogonal to the tape's line of motion, the tracks are angled across the width of the tape and all of the tape is used on just one pass.

In turn, these differences in the way data is laid out on the tape raise interesting implications for performance. The universal truth for all tape drives is that performance is utterly dependent upon the ability to keep the tape streaming across the head. Anything that interrupts the flow of data will significantly affect overall performance as the system is forced to stop and reposition the tape. When such an event occurs, the physical rather than relative speed at which the tape is moving past the head will be a more determining factor in how long it takes to reposition the tape and resume writing data.

In addition, the necessity to move the tape rapidly and stop and reverse direction in order to record on the next set of tracks places a high degree of stress on the tape media. This stress makes the tape more susceptible to wear and damage. In the case of current DLT drives, tension on the tape can be as much as 133 grams of tension on the tape, which is more than an order of magnitude greater than the Mammoth-2's servo-controlled, direct-drive dual-reel mechanism, which only exerts 12 grams of tension on the tape.

Head games

Interestingly, the 8-mm helical-scan Mammoth-2 shares more in common with HP's 4-mm DAT technology-which is also helical-scan-than it does with DLT. In the January 2000 issue, InfoStor reviewed the new generation of DDS-4 tape drives. For 4-mm DAT, DDS-4 radically improved performance with a threefold increase over the previous generation of DDS-3 drives. Exabyte has implemented many of the same types of improvements to extract a fourfold performance boost for Mammoth-2 over the first generation of Mammoth drives.

To get a threefold boost in performance in the new generation of DAT drives, HP mounted the preamplifier directly on the rotating drum. This move was made to keep the preamplifier as close to the signal source as possible in order to maintain a good signal-to-noise ratio at high transfer rates. Exabyte has adopted this scheme as well in what it calls "Power-on-rotor technology." By placing the signal-amplification circuitry for both the read and write heads on the spinning portion of the scanner, signal strength for both the Mammoth-2's read and write channels are nearly equal.

Figure 2: The Mammoth-2 drive features a number of technical advances that target the rotating scanner drum, which holds the read and write heads. The preamplifier is mounted directly onto the drum to put it as close to the signal source as possible. Grooves are precisely cut in the drum to create an air-bearing surface and a third head is added to create an air dam to maintain optimal (minimal) contact between the tape and the tape heads.
Click here to enlarge image

With the equality in signal strength, Exabyte is now able to transmit both the read and write signals through the transformers simultaneously. As a result, Mammoth-2 implements a true read-while-write operation, which obviates the need to alternate between writing data and reading to validate data. This doubles throughput and permits the Mammoth-2 to continue writing for a full rotation.

The equality of signal strength also makes it possible to move the heads closer together-90 degrees apart rather than 180-on the circumference of the drum. HP was also able to configure read and write heads at 90-degree angles; however, where the DAT drive has two read and two write heads, the Mammoth-2 drive doubled the number of heads to four read and four write heads. On each revolution of the scanner, the Mammoth-2 writes four tracks of data while simultaneously reading the previous four tracks of data. Thus, by doubling the write time and doubling the write heads, Mammoth-2 raises performance levels by a factor of four over the previous generation of Mammoth drives.

To provide better contact between the tape and the read/write heads on the drum at their high rotation speed, both HP and Exabyte have machined grooves on their respective scanners to create a negative air-bearing surface. In addition, Exabyte adds a conditioning head in front of each read and write head pair. This conditioning head acts as an air dam to reduce drag, lifting the tape before it reaches the read and write heads and insuring that the tape touches both heads with only enough force to reliably read or write data.

Another important technology that both the helical-scan drives use to maximize both the efficiency and reliability of signal detection is partial response maximum likelihood (PRML). As a recovery technique, HP starts with a Class 1 PRML recording technique and enhances the standard two-state Viterbi trellis with a six-state advanced sequence detection (ASD) scheme. Exabyte is the first to use an enhanced Class IV PRML (EPR4) technique similar to that used by many HDD manufacturers. EPR4 increases the number of times and levels at which each magnetic transition is sampled from two {0, }pm1} to three times {0, }pm1, }pm2}.

Media, however, continues to set the Mammoth family apart from all other tape drives with the use of advanced metal evaporated (AME) media. The magnetic layer of AME media is applied by evaporating the magnetic material onto the base layers in a vacuum chamber using an electron beam. This process creates two magnetic layers of pure cobalt that contain no binders or lubricants. These layers are then coated with a hard carbon protective layer and then a lubricant. The result is a tape with greater magnetic density and reduced physical thickness.

Reel results

InfoStor magazine benchmarked an Exabyte Mammoth-2 external drive sporting a low-voltage differential (LVD) Ultra2 SCSI interface. We tested the drive using a 64-bit Adaptec 29160 Ultra160 SCSI adapter in a Dell 2400 PowerEdge server in order to avoid any potential throughput bottlenecks.

To test the Mammoth-2, we ran the Nova Technica Tape Benchmark v1.0. This benchmark allocates a large block of memory from which it streams data to the device. By streaming directly from memory, the benchmark eliminates bus-bandwidth contention with other devices. The data can be streamed in block sizes of 2n kB, where n ranges from 0 to 6. Because most backup packages now use 64-kB transfers, we limited our tests to this size block.

Figure 3: The Mammoth-2 media is fabricated with an AME process that adds an extremely hard, thin coating to the magnetic layers that enhances performance, durability, and archival life. The hardness and anti-corrosive properties of the RSPC coating improves tape durability and reduces tape wear and resulting tape debris, allowing the media to achieve a 30-year archival rating.
Click here to enlarge image

The benchmark also generates two types of data stream: purely random data and data that falls into a preset frequency pattern. The purely random data is not compressible and therefore provides a baseline for native throughput. When run with compression turned on at the tape drive, an uncompressible random data stream also provides a worst-case scenario for the tape drive. The drive will attempt to compress this data-which is typical of highly compressed JPEG, ZIP, and e-mail archives-and in so doing throughput will degrade.

The patterned data stream was originally devised and calibrated using Exabyte Mammoth-1 and Quantum DLT 7000 tape drives. This data stream consistently produces a compression ratio on the order of 1.9-2.1 on earlier-generation tape devices. With this in mind, we used 3-GB data sets in all of our tests in order to reach a repeatable steady state in terms of data compression and throughput.

In all of our tests, the Mammoth-2 drive provided eye-popping results. The Nova Technica Tape Benchmark pegged the base throughput-no hardware compression-of the Mammoth-2 at 11.8 MBps. In contrast, the base throughput of the HP DDS-4 drive was pegged at 2.95 MBps. When we ran the compressible data stream with hardware compression, throughput on the Mammoth-2 soared to 22.8 MBps.

This result represents a compression factor of 1.93, which is in line with previous results with Mammoth-1; however, the Mammoth-2 uses a new adaptive lossless data compression (ALDC) algorithm that is rated as providing an average compression ratio of 2.5:1 across multiple data types. In contrast, we had earlier measured a compression boost of 2.66 x on the HP DDS-4 drive. Rather than a new algorithm, the boost in the new HP drive comes from using a firmware-based SCSI bus "intelligent disconnect," which optimizes the time to burst data into a mini-buffer used to hold data before the compression engine.

Comment and Contribute
(Maximum characters: 1200). You have
characters left.