IBM Ultrastar 146Z10
|
|
IBM managed this rapid time-to-market through the utilization of lower-capacity platters. While the flagship Cheetah 10K.6, Fujitsu's MAP3146, and the upcoming Maxtor Atlas 10k IV all utilize four 36-gigabyte disks, the 146Z10 uses 6 platters that store just under than 25 GB each. IBM specs the drives average read seek times at 4.7 milliseconds. An 8-megabyte buffer and 5 year warranty round out the package.
There's been much speculation running rampant on the merger of IBM and Hitachi's hard disk divisions into a new separate entity. According to our contacts at IBM, however, the 146Z10 remains a pure Big Blue product. The new company is not expected to merge technologies or product plans until the transition is complete.
The 146Z10 is one of the first Ultra320 SCSI drives to hit our testbed. Ultra320's most highly-touted feature, of course, is an increase in the throughput ceiling to 320 MB/sec. Keep in mind that the current transfer rate champion, the Cheetah 15K.3, pushes 76 MB/sec in its outer sectors. While figures like these threaten Ultra2 SCSI's 80 MB/sec barrier, Ultra160's 160 MB/sec limit maintains plenty of headroom. It is only in multi-drive scenarios that 160 MB/sec bottlenecks can arise. Our base drive tests occur with a single unit as the only active device on the host adapter. In such a setup, any performance advantages that Ultra320 (along with the requisite higher-bandwidth 64-bit PCI slots) would deliver over 160 are negligible. For our performance tests, we're going to take advantage of the specification's backwards compatibility and run the drive in Ultra160 mode off of our current host adapter. Bear in mind that improved bandwidth is only one of the benefits that Ultra320 delivers. A host of improvements in protocol and error correction should elevate data integrity and device interoperability to new levels.
