era@niwot.scd.ucar.edu (04/07/90)
This is a summary of information I received from a sun-spots request I made about IPI-2 interface disks for Sun a while back. I was inquiring about what Sun was actually selling under their 706A/707A options in the price book, and what 3rd party sources were offering. The line of disks Sun is currently selling for 4/490 systems are from Imprimis Technology of Minnetonka, MN (formerly owned by CDC; now owned by Seagate). The Imprimis "Sabre" line of 8" disks with data transfer thru a single head at a time, consists of the 9720-368, 9720-500, 9720-736, 9720-850, and 9720-1230, where the number after the dash indicates the number of megabytes, unformatted. These drives have either 7 or 9 platters, TPI ranging from 960 to 1289, BPI ranging from 14981 to 25264, and average seek time of 18 or 16 ms. Sun's 706A option from their 1 Jan. 1990 catalog, is a 9720-1230. (IPI-interface disks are also being offered by Fujitsu, Hitachi, and NEC, but I have no information that Sun will accept those on their maintenance contracts. Besides, you want to Buy American, right? :-) ) For increased capacity, there is a new double-density model, the 9720-2500. One contact claims that this drive has a slightly better average seek (i.e. < 16ms). However, we have not been able to get any literature on it and have not found any vendors who are stocking it yet. For increased transfer rate, there is a dual-head-transfer model, the Sabre 2HP 1150, which will transfer data simultaneously thru 2 heads at a time. This is *not* a dual-actuator drive; all heads are tied to one actuator. Although one contact says his Sun salesman told him that Sun's 707A *is* this drive, *our* salesman has told us that the 707A is just the same as 706A with mounting hardware added, and that's what the 1 Jan. 1990 catalog shows; price is the same. Vendors seem to use slightly different terminology for disks in the Sabre series, which apparently has to do with the type of interface (SCSI, SMD, or IPI) added to the drive. For instance, one vendor offers the 9720-1230 with IPI interface as a CDC-97209-12G, where IPI is specified by the '9' following "9720" (0=SMD,1=SCSI,9=IPI). One vendor claimed the 4-digit CDC designations are being changed to 5 digits because of the Seagate acquisition. Format.dat 1.26 of 89/09/06, from a 490, shows Sun support of the following Imprimis drives on either the SMD-4 or ISP-80 (Sun IPI-2) controllers: disk_type = "CDC EMD 9720" \ : ctlr = XD7053 \ : ncyl = 1147 : acyl = 2 : pcyl = 1217 : nhead = 10 : nsect = 48 \ : rpm = 3600 : bpt = 30240 : bps = 613 disk_type = "CDC 9720-850" \ : ctlr = XD7053 \ : ncyl = 1358 : acyl = 2 : pcyl = 1360 : nhead = 15 : nsect = 66 \ : rpm = 3600 : bpt = 41088 : bps = 610 disk_type = "CDC IPI 9720" \ : ctlr = "ISP-80" \ : ncyl = 1631 : acyl = 1 : pcyl = 1633 : nhead = 15 : nsect = 82 \ : rpm = 3600 : bpt = 50400 : skew = 0 : precomp = 0 The 3 disks shown above are, respectively, the Sabre 9720-368, 9720-850, and 9720-1230. Sun has tested the 2HP 1150 model internally, but apparently is not offering them because they aren't currently available in quantity. One contact, a large/favored Sun customer, asked Sun for this drive but was turned down; however, another contact indicated that his Sun salesman let him order one of these. Our Sun salesman had the following to say about performance of the 2HP model: |The work that we have done shows preliminary results that |sequential operations are about twice as fast with 6MB/s |drives versus 3 MB/s drives. Dual heads also seem to improve |the random case by 20% to 30% since the lay out of the data |essentially doubles the logical track size. This automatically |forces a greater degree of sequentiality even in the random case. Solbourne also has an IPI disk program. The manager of their disk program claimed that they would be offering IPI "within a few months", and had the following to say about the 2HP drives: |Yes, I'd agree with the statement that the position/placement of |data is transparent to the host. It is all handled within the |drive and controller. |Yes, we have two or three 2HP drives in-house now from Seagate. |We are still tuning the IPI-controller/host interation so the |numbers we're seeing now are just preliminary. We are confident |we'll attain off each drive the physical data rate of 6 Mbyte/sec, |a raw data rate of 4.5 Mbyte/sec and a file system data rate of |4 Mbyte/sec. The steep fall-off in physical to raw data rates is |due to the long data headers IPI defines for each 512 byte data |block. I think it's on the order of 128 byte per block compared |with 90 bytes per block under SMD. The Imprimis sales literature for the 2HP confirms the nature of the interface: "The splitting of the data on a write and the combination of data on a read is done entirely by the disk drive interface and is transparent to the disk controller and/or adapter." That is, the drive appears to the system to have half as many tracks per cylinder, with tracks that are twice as long as for the single-head-transfer model. As far as availability of the 2HP model, I didn't come across anyone who actually had one at the time my request was written. A number of vendors claimed that they couldn't get them. However, one contact said that Peripheral Designs (see below) claims they can get them, for a price of ~$11500. One other contact claimed that vendors have offered him the 97229, although he didn't say which vendors. Since one contact claimed that Sun had signed an exclusive deal with Imprimis, it may be that the 2HP drive is only available from sources who had contracts for it that predate Sun's contract. Several contacts have purchased the 9720-1230 drives from 3rd party suppliers and formatted/run them without problems, so 3rd party sources can save you quite a bit, esp. if you don't qualify for Sun's category B discount. Following are the 3rd party sources we've identified so far. The lowest prices we've heard of for the 97209-1230 and 97229-1150 have been ~$6800 and ~$11500, respectively. Andataco 800-334-9191 Aviv 617-933-1165 Consan; Eden Prairie, MN 612-949-0053 Cranel 614-433-0045 Falcon Systems, Sacramento 916-344-1221 Microtek ? National Peripherals ? Peripheral Designs 404-263-0067 Peripheral Devices, Chicago 708-894-9050 R2D Distributing 303-799-9292 Rorke Data [wholesale only, I believe] 612-829-0988 CAVEAT: when you get that best price, make sure it includes the front panel (controls), cables, and mounting hardware. Some vendors don't seem to have, or even be aware of, what kind of hardware is required to mount these drives in the new Sun Data Center Server cabinet. One vendor has told us that the 19" rack hardware will work in the Server cabinet, but we haven't been able to verify that. As far as controllers, we have not considered using anything but Sun's ISP-80, which was developed internally at Sun. Those who don't use Sun support may be interested to know that Xylogics will have their IPI-2 controller available for delivery from their distributors in early May, in both single and dual channel models. (The dual-channel model acts like two controllers operating in parallel, saving you a slot.) I don't know who all their distributors are, but I believe Falcon is one. Disclaimer: All of the above, or none of the above, may be true. Ed Arnold * NCAR * POB 3000, Boulder, CO 80307-3000 * 303-497-1253(w) era@ncar.ucar.edu [128.117.64.4] * era@ncario.bitnet * era@ncar.uucp "Life is kind of like a B-grade movie; you don't want to cut it short, but you don't want to roll it again, either." --Ted Turner
kwthomas@nsslsun.gcn.uoknor.edu (Kevin W. Thomas) (04/14/90)
In v9n115, era@niwot.scd.uar.edu writes: >For increased capacity, there is a new double-density model, the >9720-2500. One contact claims that this drive has a slightly better >average seek (i.e. < 16ms). However, we have not been able to get any >literature on it and have not found any vendors who are stocking it yet. Try "Computer Upgrade Corporation" in Aneheim, California. They can be reached at 714-630-3457. We received a quote, including the 7053 controller for under $13,000. Kevin W. Thomas National Severe Storms Laboratory Norman, Oklahoma