Seagate Cheetah 15K.4
Note: Since the publication of this review, this drive has been retested under Testbed4, a newer hardware/software/benchmark platform. Please see this article for updated results. This review remains for reference purposes only.
|
|
2001's Cheetah X15-36LP further tweaked the X15's design with yet lower seek times and a higher 36-gigabyte capacity. It also heralded the maturity of fluid dynamic bearing motors, a feature that previously found a home only in less robust ATA designs. The integration of FDB motors finally meant that whisper-quiet idle operation and ultra-high rotation speeds were no longer mutually exclusive.
Seagate's 2002 offering, the impressive Cheetah 15K.3, further refined noise and heat output while delivering an eminently-usable 73 gigabytes of storage. By this time, however, competition arrived from the likes of Fujitsu, Hitachi, and Maxtor. Though it retained a proud first-mover heritage, the 15K.3 certainly received a run for its money when it came to sheer performance.
Enterprise drive manufacturers took a breather in 2003 as the realities of both media and head limitations gripped the industry. While ATA drives inched forward here and there, SCSI drive families, ostensibly a level above their consumer-class brothers in reliability, rested where they were. This year, however, finally saw the announcement of next-generation product from all major players.
The Cheetah 15K.4 packs over 36 gigabytes of data on each of its four miniscule 2.6 inch platters to expand flagship capacity up to 147 GB. Seagate's specs cite a 3.3 millisecond seek time, shaving 0.3 ms off of the 15K.3's claim. Buffer size remains an industry-standard eight megabytes.
With the 15K.4, Seagate sees its competitive advantage resting more than ever in reliability and rock-solid operation. The firm has bumped the drive's Mean Time Between Failure spec from 1.2 million to 1.4 million hours and claims the 15K.4 is the quietest and coolest 15K Cheetah yet.
Seagate's SeaTools Enterprise utility provides diagnostic and configuration functions for the manufacturer's SCSI drives. One particular setting permits the user to modify buffer segmentation settings between static allocation (hence allowing data read in advance to remain longer) and dynamic allocation (allowing the size of buffer segments to change as loads and patterns vary). The former paradigm achieves higher buffer hit rates in localized patterns (single-user applications) while the latter tends to lend advantage to highly-random accesses (multi-user applications). In the past, these modes did not drastically change the performance of Seagate's offerings. With next-generation drives such as the Cheetah 15K.4, 10K.7, and Savvio 10K.1, however, significant differences surface. For the purposes of this review, we have run tests on the 15K.4 in both Desktop (static segment allocation with the drive's default of 32 segments) and Server (dynamic allocation) modes.
The Cheetah 15K.4 aims to maintain Seagate's share of the market servicing databases, data warehousing/mining, transaction servers, and other applications that feature highly-random accesses under heavy loads. In the following tests, the Cheetah 15K.4 will be compared against the following drives for the following reasons:
| Fujitsu MAS3735 (73 GB) | Manufacturer's previous-generation unit |
| Hitach Ultrastar 15K73 (73 GB) | Previous-generation competing unit |
| Maxtor Atlas 15K (73 GB) | Previous-generation competing unit |
| Seagate Cheetah 15K.3 (73 GB) | Manufacturer's previous-generation unit |