| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 4696.1 | | TENNIS::KAM | Kam WWSE 714/261.4133 DTN/535.4133 IVO | Sat Jul 06 1996 13:10 | 33 |
| From www.storage.digital.com, the page has this advertisement:
Digital's New EZ3x and EZ6x Solid State Disks
...for mission critical applications
From VTX IR | SU | Keyword: solid state disk
1 24-Apr-96 Product and Service Retirements SU3125
2 24-Apr-95 Product and Service Retirements SU1BB6
3 10-Oct-94 HSC High Performance Software V8.4 Expands SCSI SU1AUB
Support and Adds Jukeb
4 08-Sep-94 KZPAA PCI-to-SCSI Host Bus Adapter SU1AON
5 11-Jul-94 StorageWorks HSD30 Array Controller SU1ADN
6 02-May-94 StorageWorks HSZ40 Array Controller SU14A7
7 07-Mar-94 The StorageWorks Family Announces New BA350 Shelf SU13Z3
Upgrades
8 21-Feb-94 EZ58R SCSI Solid State Disk Now Available SU13WC
9 29-Nov-93 EZ5x Additional Interconnect, System, and Operating SU13FP
System Support
10 04-Oct-93 EZ5x SCSI-2 Solid State Disks Now Shipping SU136V
11 26-Jul-93 New StorageWorks Solutions for Digital's VAX and SU12QI
Alpha AXP Systems
12 26-Jul-93 EZ5x High Performance SCSI-2 Solid State Disks SU12QM
13 26-Jul-93 Storage Resources SU12QT
o #EC-N0004-45, "Solid State Disks," a comprehensive handbook
I can't help with the RAMDISK article because I can't get access to the
QUORUM information on TIMA.
Regards,
|
| 4696.2 | DECRAM = RAMdisk | JOBURG::HARRIS | | Sun Jul 07 1996 06:14 | 5 |
| Finally found some info ....DECram was the keyword that did the trick!
Now All I need is a comparison whitepaper!
Ivan
|
| 4696.3 | | SSDEVO::PARRIS | Keith, Digital Consultant | Sun Jul 07 1996 20:16 | 6 |
| For most customers, the basic question is: Will the data be modified? If not,
use DECram, because you don't care if the data goes away if the power fails or
the node crashes; you can always reload it on reboot. If the data will be
modified, use a solid-state disk for its non-volatility. Performance-wise, the
DECram will win hands-down, because it uses main memory and avoids the latency
of going out over some I/O bus to the solid-state disk.
|
| 4696.4 | "Ram Disk is *not* an installation procedure" | JGODCL::APETERS | Let's make it happen! | Mon Jul 08 1996 04:04 | 1 |
| couldn't resist ;-)
|
| 4696.5 | what will it cost me to lose these data? | WRKSYS::HOUSE | Kenny - PKO3-1/N75 - DTN 223-6720 | Mon Jul 08 1996 09:03 | 18 |
| re .3 ...
> For most customers, the basic question is: Will the data be modified?
--------
> If not, use DECram, because you don't care if the data goes away if the
> power fails or the node crashes; you can always reload it on reboot.
Actually, the question probably should be "are the data on the RAM disk
recoverable at a reasonable cost?" If you can reload raw data and
rerun the job in a short time, this may be acceptable to most
customers. I wouldn't do data entry or data capture to a RAM disk, nor
would I run a month-long job (that my boss was waiting for) without
some sort of non-volatile checkpointing.
As Keith pointed out, it does make sense to keep data in memory for
performance.
-- Kenny House
|
| 4696.6 | Ramdisk: I like that name | INDYX::ram | Ram Rao, PBPGINFWMY | Mon Jul 08 1996 11:51 | 5 |
| Re: Ramdisk Product and Performance specs wanted urge
I am honored that Digital would choose to name a product after me :-)
Ram
|