pan@propress.com (Philip A. Naecker) (11/18/89)
In article <35985@ames.arc.nasa.gov>, lamaster@ames.arc.nasa.gov (Hugh LaMaster) writes: ... > 1) Why are SCSI disk subsystems *so* slow. On sequential reads, ~300 KB/sec. > is typical. The filesystems and all hardware interfaces should be able to > easily sustain 3X as much on sequential reads of large, unfragmented files. > (Which SCSI disk subsystems you ask? I would rather turn it around, and say, > are there any exceptions to the above sweeping generalization? :-) I use > SMD disks as a baseline...) Please define your terms more precisely. I think what you mean for "slow" is "have such low transfer rates". This is not always the meaning of "slow". In particular, in many general computing (timesharing, software development, workstations, etc.) or OLTP environments, I would submit that access time (both seek and rotational latency) is much more important than transfer rate. For example, I refer you to the October issue of DEC Professional Magazine, which has the third part in an excellent series by Ken Bates of DEC, showing that most VMS timesharing IO's are quite small and involve no seeking at all. You might also try the latest issue of Computing Systems, a journal from USENIX, which has an excellent article on the subject of UNIX disk IO and includes some relevant statistics. > > 5) (Not really a comp.arch question, but related to the above - aside: > Does anyone make synchronous SCSI disks which really perform?) If transfer rate is what you seek, you should take a look at disk arrays. There are many variants to such devices, but most will gang several disks into a single large logical volume that is rotationally synchronized. The effect is to make many more heads active reading data at a time, resulting in much higher transfer rates. Striping drivers are a software solution that provides similar results. There are now several manufacturers of these wonderful devices. Try, for instance, EXSYS and Micro Technology - there are others, too. _______________________________________________________________________________ Philip A. Naecker Consulting Software Engineer Internet: pan@propress.com Suite 101 uunet!prowest!pan 1010 East Union Street Voice: +1 818 577 4820 Pasadena, CA 91106-1756 FAX: +1 818 577 0073 Also: Technology Editor DEC Professional Magazine _______________________________________________________________________________