Richard Gambrell ([log in to unmask]) wrote:
: I agree. But we love MPE's reliability and dependablity - if we can
: have it with true OPEN, "plug and play" type standards - then great. If not,
: HP ought to be clearer about it - perhaps calling it HP-SCSI.
: I hope someone from HP can clarify what we "gain" by the "special"
: nature of SCSI disk drives for MPE. Still, I like this situation better
: than a unique, (e.g. HPFL) solution - we do gain from the wider experience
: everyone has with SCSI, even if we do have to buy HP MPE "special" SCSI
drives
: to work with an HP3000.
We get to hear this quite a bit so...
** This is what the HP customer gets when he spends the bigger dollars: **
1. Performance Optimization (i.e. custom 'tuned' firmware for HP systems)
HP mechanisms are optimized for high performance with all HP CPU platforms.
Very significant, noticeable disk I/O performance increases are seen.
40% improvements, though not guaranteed, are not uncommon even over HP's
distributor mechanisms! Test data shows a 17% improvement with a C3027U
(1Gb HP workstations) over a C2247 (1Gb distributor) utilizing RANDOM I/O,
75% writes, with 4 single ended disks, 8k transfers, 50Mb xfers on a
720 system. Using 25% writes a 6% improvement is seen.
115% improvement is seen with SEQUENTIAL I/O's, 75% writes, with 1 single
ended disk, raw I/O, 100Mb xfers on a 720. Using 25% writes an 18%
improvement is seen.
Keep in mind that an 18% improvement is about the same improvement that is
seen from one disk generation to the next (about a 1 year time advantage).
2. Functionally Unique Features, Power Fail Support for S800/900
Drives for the S800/900 have a unique feature for 'power fail support'
technically called Sector Atomicity. HP 6000 Series Disk Drives have
this feature. HP distributor AND competitors disks do NOT have sector
atomicity. A patent on the feature has been applied for.
This feature ensures that no incomplete writes to the disk will occur if
power is lost. This eliminates the possibility that incompletely written
data, i.e. corrupt data, is stored, and ensures that only complete,
correct data is on the disk.
3. Extensive Product Integration and Compatibility Testing
HP goes to great lengths to ensure our products will work together with our
systems. This involves literally thousands of hours of mechanism test
time. HP does this compatibility testing between its systems and disks
because IT DOES FIND PROBLEMS. Don't accept arguments that open systems
components will always work together. The amount of testing HP does is
one of the MAIN reasons that HP products are so reliable.
4. Guaranteed Support
Below is a partial list of test which HP does to ensure that our products
work:
* Disk drive configuration testing
* Regression testing (Do commands, options, transfer sizes, abort
sequences work)
* Power fail testing (What happens when the power fails?)
* ESD testing (Are these disks and HP CPUs immune to electrostatic
discharge?)
* Fault testing (When faults occur, does the system handle them?)
* Beta testing (Do the products work in representative customer
environments?)
* Diagnostic testing (Do the diagnostics work correctly?)
* Performance testing (Is the disk as fast as it should be?)
* Boot/Root functionality (Do the disks function as boot/root devices?)
* EMI/RFI testing (Does the CPU, with HP peripherals meet FCC emission
standards?)
EVERY one of these tests has produced corner case problems which HP has
fixed prior to customer shipments.
Be sure that the customer understands what support and testing the
distributor provides prior to purchase decision.
There are actual cases where 3rd party mechanisms have failed to work
when the customer rolled to the next revision of the operating system.
This is not the case for HP supported mechanisms.
5. HP Field Support
6. HP Sales Support
7. HP Response Center Support
8. HP Single Vendor Solution
9. HP Parts Stocking
10. HP Escalation Management
The customer must evaluate for himself what it will cost him IF he does have
a failure. If the customer has MISSION CRITICAL data on his disks, it would
probably be relatively easy to justify the added expense. If a failure is
only going to cause a one day delay in a development project, it may be
difficult to justify the added expense.
Regards,
Phil Haseltine
*** I do an array of disk support. ***
Oh no! Now I have two e-mail sites!
[log in to unmask]
[log in to unmask]
This response does not represent the official position of, or statement by,
the Hewlett-Packard Company. The above data is provided for informational
purposes only. It is supplied without warranty of any kind.
If it's broke, we fix it.
If it isn't broke, we enhance it.
If you find it acceptable, then your
expectations are probably too low....
: --
: -- - - - Speaking for myself and not necessarily anybody else - - - - - -
: Richard Gambrell | Internet: [log in to unmask]
: Mgr. Tech. Services | POT: 504-483-7454 FAX: 504-482-1561
: Xavier University of LA | Smail: 7325 Palmetto, New Orleans, LA 70125
|