Subject: comp.arch.storage FAQ 1/2
Date: 29 Nov 1995 23:18:14 GMT
Expires: Sun, 31 Dec 1995 00:00:00 GMT
        tape, RAID, file systems, hierarchical storage management

Archive-name: arch-storage/part1
Version: $Header: /jpn/rdv/comp-arch-storage/RCS/FAQ-1.draft,v 1.22 95/11/29 15:04:56 rdv Exp $
Posting-Frequency: monthly

Rod Van Meter, Joe Stith, and the gang on comp.arch.storage<BR>
rdv@isi.edu or rdv@alumni.caltech.edu

Apologies for being late getting it posted! PLEASE, if you have
additional information for me to include here, send it to me. I think
I'm caught up on all of the additions people have sent in the last
couple of months.

Information on disk, tape, MO, RAID and SSD can be found in part 1 of
the FAQ.  Part 2 covers file systems, hierarchical storage management,
backup software, robotics, benchmarking, MTBF and miscellaneous
topics.


        Many items merely identified, not described.

        Last updated: 1995/08/16

        Most recent changes:
..lots of (mostly small) changes!
..PC removables reference started (ZIP, Jazz, SyQuest)
..Sony 3.5" MO
..INLINE RES-400 tape drive
..MountainGate
..Sony ID-1 updated
.    (part 2)
..DATMAN
..HSM evaluation
..MountainGate
..network-attached peripherals

        comp.arch.storage:
                Storage system issues, both software and hardware


1.     Editor's Note

2.     Disclaimer

3.     Original Editor's notes

4.     Truly Frequently Asked Questions

5.     Tape
5.1.   Cartridge vs Cassette
5.2.   Longitudinal
5.3.   Serpentine
5.4.   Helical Scan
5.5.   9-track {brief}
5.6.   3480/3490/3490E {brief}
5.6.1. New IBM Tape (NTP) 3590 {New}
5.7.   QIC {brief}
5.7.1. Travan {new}
5.8.   4mm {brief}
5.9.   8mm {brief}
5.9.1. Mammoth {Brief,New}
5.10.  DEC/DLT {full}
5.10.0.1. DLT4000 (From DEC)
5.10.0.2. DLT2700 (from DEC)
5.10.0.3. DLT2000 (from DEC)
5.11.  MountainGate (was Metrum) VHS {brief}
5.12.  VCR VHS
5.13.  19MM (D1 and D2) {Brief}
5.14.  ID-1
5.15.  D-2
5.16.  StorageTek Helical {Brief}
5.17.  Optical {new}
5.18.  D-6 {brief}
5.19.  D-3 {brief}
5.20.  INLINE RES-400 {Brief,New}

6.     Disk
6.1.   CAV, ZCAV and CLV
6.2.   Optical {Brief}
6.2.1. CD-ROM
6.2.1.1. Next-Generation CD-ROM {New}
6.2.2. WORM {brief}
6.2.2.1. Sony {brief,new}
6.2.3. Erasable
6.2.3.1. Magneto-Optical Physics
6.2.3.2. Magneto-optical, 5.25-inch
6.2.3.3. Magneto-optical ZCAV, 5.25-inch
6.2.3.4. HP Corsair {Brief, New}
6.2.3.5. Maxoptix T4-1300
6.2.3.6. Asaca HSMO {Brief}
6.2.3.7. Other Multi-beam MO {None}
6.2.3.8. 3.5-inch MO {Brief, New}
6.2.3.8. Sony {Brief, New}
6.2.3.9. Nikon 12-inch MO {Brief}
6.2.3.10. Sony 12-inch MO {Brief,New}
6.2.3.11. NEC 12-inch MO {Brief,New}
6.2.4. Electron-Trapping {None}
6.2.5. Dual Function {Brief, New}
6.2.5.1. Panasonic/Toray {Brief, New}
6.2.5.2. IBM {None}
6.3.   Magnetic
6.3.1. 5.25-inch
6.3.1.1. Seagate
6.3.2. 3.5-inch
6.3.2.1. IBM
6.3.3. Bernouli {None}
6.3.4. Floptical {Brief}
6.3.5. PC Removables {Brief, New}
6.3.5.1. SyQuest EZ135 {Brief, New}
6.3.5.2. Iomega Zip {Brief, New}
6.3.6. Mainframe {Brief}
6.4.   Other
6.4.1. Holographic Storage Products {Brief}

7.     RAID {Full}
7.1.   RAID Levels
7.2.   RAID-6
7.3.   John O'Brien and RAID-7
7.4.   RAID Papers
7.5.   R-Squared {Brief}
7.6.   Sun {Brief}
7.7.   the RAIDbook {Brief,New}
7.8.   Software Striping {Brief}
7.9.   RAID Vendors {New}

8.     Solid State Disk (SSD) Frequently Asked Questions (FAQ)

9.     Other Devices
9.1.   SyQuest Removable Cartridge hard Drives
9.2.   Kalok removable cartridge hard drives

10.    RAIT (Redundant Arrays of Inexpensive Tape)

11.    RAOT (Redundant Arrays of Other Things :-)


------------------------------
Subject: [1] Editor's Note
From: Editor's Note

I took over the maintenance of the C.A.S FAQ from its originator, Joe
Stith (stith@fnal.gov, Fermi National Accelerator Laboratory, at the
time), in July '94. I have begun additions and edits of my own. It
should be available on the various FAQ servers. Apologies for the
under-construction formatting and lack of better referencing.

I will add as my expertise and time allows and will include
submissions sent to me (Rod Van Meter, rdv@alumni.caltech.edu) and
information put into the newsgroup by others.  Yes, some of the
submittals are from vendors (including me, see the disclaimer).  If
you post to the newsgroup and find yourself quoted in this FAQ but
wish to be removed, please let me know.

There is more information in my (still primitive) WWW version of the
FAQ, including more commercial info. It is temporarily available at
http://alumni.caltech.edu/~rdv/comp-arch-storage/FAQ-1.html, but this
is probably not a good permanent home for it (volunteers?). I have
also started working on a Japanese translation available there (only
about 10% complete). There is also a good group of FAQs (including
this one) stored at http://www.cis.ohio-state.edu/hypertext/faq/;
prettier and easier to use but with only the info I actually post.

The size of these FAQs is getting out of hand; I'm open to suggestions
on material that doesn't really belong in a FAQ or areas where I'm
simply too verbose (I may reduce the DLT info since there is now a
separate FAQ).

SHMO: when you see this, it means "Somebody Help Me Out!" I'm actively
soliciting information on this topic.

See the copyright notice near the end.

------------------------------
Subject: [2] Disclaimer
From: Disclaimer

I used to work for ASACA, which makes Metrum's (now Mountain Gate)
robotics and makes a 12 MB/sec. magneto-optical disk drive, and
resells AMASS in Japan. I now work at USC's Information Sciences
Institute working on, essentially, network-attached peripherals and
high-speed networking.  This information is included so you can
identify my bias. Obviously the things that I know the most about are
the best-represented. I attempt to be as impartial as possible; if you
have complaints about my fairness, let me know. None of this should in
any way be construed as the official opinions of ASACA, USC/ISI or
Caltech, and may not even represent MY opinions.

------------------------------
Subject: [3] Original Editor's notes
From: Original Editor's notes

I believe a reference would be useful and I am willing to pull
it together...

I have included the original call for votes and will go through that 
list for other ideas to include.

I will also format this for sending to news.answers, misc.answers, 
and comp.answers.

As I am writing this, "{None}" indicates I have not written anything 
for it yet, {Full} indicates it is OK, while {Brief} indicates somewhere in 
the middle.

{New} indicates some new information has been provided.

=========DISCLAIMER=====

        This information is believed to be reasonably accurate although I do 
not verify every submission.  No legal liability is assumed.  See full 
disclaimer at end.

------------------------------
Subject: [4] Truly Frequently Asked Questions
From: Truly Frequently Asked Questions


Also see the miscellaneous section near the end of part 2 of the FAQ.

* How do I connect X to Y?

Try asking in the appropriate comp.periphs.x or comp.sys.y groups.

* What about the jumper settings for my Yoyodyne 4000?

Try asking in the appropriate comp.periphs.yoyodyne group.  This group
attempts to keep the discussion at a higher level.

* Can somebody recommend a PC or Unix backup package? (SHMO)

The Aug. '94 issue of PC World has an article covering PC tape
drives (mostly QIC), and covers backup software to some extent as
well. Try asking this question in the PC-related newsgroups.

I've started adding a list of backup software to part 2 of this FAQ,
though it's not very complete yet.

* What about DAT/DLT/8mm/...?

Covered in the sections on tape drives & media. There is also a
DLT-specific FAQ maintained by a guy from DEC posted occassionally
here.

* Does anybody have the phone number of...?
.
There's a long list currently posted to: comp.sys.ibm.pc.hardware.misc
(inside FAQ) and in biz.zeos.general.

For the WWW-enabled, check out http://www.cmpcmm.com/cc/.

* What about RAID7?

See the description under RAID arrays; NOT a popular topic around
here.

* MTBF?

This topic is responsible for generating the most heat in this group.
There are a few sentences about it in part 2 of the FAQ, but I'm not
qualified to write this section (and/or not willing to suffer the
public humiliation :-).

------------------------------
Subject: [5] Tape
From: Tape
(No compression added in calculations unless specified otherwise)

------------------------------
Subject: [5.1] Cartridge vs Cassette
From: Tape
Cartridge has only one reel (i.e.  3480, DLT).  Cassettes have two
reels (i.e., 8mm, 4mm, 19mm, VHS) and may not need to be rewound to
dismount.

------------------------------
Subject: [5.2] Longitudinal
From: Tape
have heads that write bit streams that are parallel to the edge of 
the tape.

------------------------------
Subject: [5.3] Serpentine
From: Tape
are longitudinal that write the full length of the tape, then turn 
around and write the length of the tape in the opposite direction with the 
heads in a slightly different position.  This process may repeat many times.

------------------------------
Subject: [5.4] Helical Scan
From: Tape
are like your VCR with a rotating head mounted at an angle writing 
"swipes" at an angle not parallel to the edge of the tape.  The tape is moved 
only slightly between swipes.  Two or so longitudinal tracks may also be used 
for fast positioning purposes.

------------------------------
Subject: [5.5] 9-track {brief}
From: Tape
The old 2400 foot round reel of tape written at 800, 1600, or 6250 BPI.

------------------------------
Subject: [5.6] 3480/3490/3490E {brief}
From: Tape

        Square cartridge:
..serpentine
                IBM 3480 - 18 track - 200 MB/cartridge, 1984
                IBM 3490 - Smaller packaging, IDRC standard, 1989
                        reports that it is a 3480 in ID string
                IBM 3490E - 36 track, double length - 800 MB/cartridge 
..native, 1992
        .IDRC data compression is also available.

These drives have traditionally come from the mainframe vendors --
IBM, Fujitsu, Storage Tek, etc. They were originally very large
objects, with vacuum columns and mainframe interfaces, the size of
large filing cabinets.

However, recently they have become available in smaller packages, 19"
rack mount or table top, and with interfaces such as SCSI.

Here's one from last year that I recently dug out of some old mail:

IBM has just made generally available the following:
  3490E Model E SCSI tape drive
  fast/wide differential
  IDRC compression
  3 MB/sec at the tape head, 6.5 MB/sec with compression
  approx 800 MB tape capacity with out compression, 2.4 GB with a 3:1 comp
  7 tape CSL(cartridge stack loader)
  Desk top and rack mount
  List $27,000

wener@vnet.ibm.com, 94/4/11

There's also a standalone drive with a small autoloader from a company
called Overland Data. Their L490e is a win because it reads and writes
both 18 and 36 track tapes. At $20K it's reasonably priced. For the
www-enabled, see http://www.ovrland.com/~odisales.
(tdowty@ovrland.com, 94/12/21)

------------------------------
Subject: [5.6.1] New IBM Tape (NTP) 3590 {New}
From: Tape

Unannounced product from IBM that has been demonstrated at mass
storage conferences. Same form factor as 3480, serpentine. 9 MB/sec.,
10GB/cart. Bare drive should retail ~$35K, 1Q95. Two SCSI interfaces?
Robotics also planned. (garyblk@aera.com, 09/30/94)

NTP has been released (95/4/10) as the 3590 tape subsystem with the
Magstar tape drive. The press release didn't mention price,
availability, or interfaces (though it did say attachable to Suns,
RS-6k, etc.).

list price for the 3590 drive is approximately $43.5K;
list for media is approximately $50;
list for 3590 drive + 10 cartridge stacker is approx $48K (95/9/18,
andrew@research.att.com)

IBM Storage has an excellent WWW page at
http://www-i.almaden.ibm.com/storage/storage.htm.

------------------------------
Subject: [5.7] QIC {brief}
From: Tape
        Quarter Inch Cartridge.  Primarily low-end (ie, PC) available from 
multiple vendors but 3M is pushing it into midrange market with estimates of 
100 Gigabytes per cartridge by 1999. Two-reel cassette.

In Dec '94, 3M, Sony, HP and others announced a new format, 2.3 times
the capacity, available mid-1995. The cartridge and drive mechanics
apparently change. It will be .315 inch tape instead of .25, with 750
feet of media in a cartridge. (rdv, 94/12/16)

------------------------------
Subject: [5.7.1] Travan {new}
From: Tape

Travan is a 3M trademarked name for a new family of data minicartridge
tape media.  Travan provides additional capacity via longer
length(750' / 229 meters) and greater width(.315" / 8mm) tape then
previous generations of data minicartridge.  Drives that can
accommodate Travan cartridges must have a mechanism that can handle
wider tape (.315" / 8mm vs. earlier data minicartridge that are
.25"/6.2mm) and be able to accept the slightly larger Travan
cartridge.  All major drive manufactures have modified existing drives
or designed new mechanisms to accommodate the Travan cartridge.
Today's Travan capable minicartridge tape drive maintains the industry
standard QIC formats. These drives can write and read smaller capacity
data minicartridge tapes of the same QIC format family(See below).

Travan TR1 capable drive functionality
Cartridge .Format..Capability.Native Capacity
DC2120..QIC-80..W/R..120MB
DC2120XL.QIC-80..W/R..170MB
QW5122F.QIC-80..W/R..208MB
TR1(Travan).QIC-80..W/R..400MB

Travan TR3 capable drive functionality
Cartridge .Format..Capability.Native Capacity
DC3010XL.QIC3010.              W/R..346MB
QW3010XLF.QIC3010.              W/R..425MB
DC3020XL.QIC3020.              W/R..692MB
QW3020XLF.QIC3020.              W/R..850MB
TR3(Travan).QIC3020.              W/R..1600MB

Note

1. TR3 capable drives can read only all combinations of QIC-80 formatted
tapes
2. 3M has developed a TR2  800MB QIC3010 format tape but has not gone to
market with it at this time
3. There will be TR4 and greater media products in the future. These
products will offer greater capacity and performance.

(thanks to mike.lakowicz@conner.com, aug 16 1995)

------------------------------
Subject: [5.8] 4mm {brief}
From: Tape

Multiple vendors.  Initially for home audio market.  Original product
held 1.3 GB on one 60 meter tape at about 180 KB/second.  Search
speeds run about 200 times nominal speed.  DDS (Digital Data Storage)
format has overtaken the DATA/DAT format.  Two-reel cassette.

http://www.dmo.hp.com/tape/_cpb0002.htm gives HP DDS drive information
http://www.dmo.hp.com/tape/c1533a.htm gives specs for the DDS2 C1533A

------------------------------
Subject: [5.9] 8mm {brief}
From: Tape

Sony developed transport designed initially for home video market.
Exabyte has U.S. rights with Kubota qualifying as a second source when
needed.

EXB 8200 model holds 2.3 Gigabytes per tape at 220 KB/sec

EXB 8500 model holds 5 Gigabytes per tape at 500 KB/sec

Search speeds of the 8200 is dismal, but is significantly improved in
later models.  Compression and half-height (standard is full-height
5.25-inch) features are also available.

------------------------------
Subject: [5.9.1] Mammoth {Brief,New}
From: Tape

Mammoth is the new Exabyte drive. It reportedly holds 20 GB
uncompressed per cartridge, with a transfer rate of 3 MB/s. Exabyte
has been touting this drive since at least April '93. It is finally
due out "soon". (rdv,95/1/13)

------------------------------
Subject: [5.10] DEC/DLT {full}
From: Tape

See http://www.quantum.com/support/bulletins/tape.html for info on
current issues concerning DLT.

Also see the DLT FAQ, maintained by Larry Kaplan, lkaplan@tdh.qntm.com.


Digital Linear Tape (DLT)
TZ87 (DLT2000) - 10 GB native per cartridge
See also robotics (DLT2700 is 7 tape library)
Ref: Digital's Customer Update, March 14, 1994
Serpentine recording.

Developed from DEC's TK50 & TK70 technology. The unit that developed
this was sold to Quantum.

DLT is the new tape technology getting the most air time around here.
There is also a DLT-specific FAQ maintained by a guy from Quantum.

From the newsgroup:

Tape uses a special hook for load/unload mechanism.

DEC is the initial vendor, but other vendors are re-selling them 
(sort of like TTI's reselling of the Exabyte 8mm tape drive).

Transfer rate of 2.5 MB/sec, but that assumes 2:1 compression, so it
is 1.25 MB/Sec native.  Likewise the 20 GB cartridge is 10 GB native.

DLT4000 ($2K upgrade from DLT2000) soon (9/94?)  Double the capacity.
Some agreement with Cypher. Still not shipping in quantity, 1/95.

Can be used on NON-DEC systems (standard SCSI interface).

One report of a batch of tapes that were "too wide".

>michelotto@mvxpd5.pd.infn.it (Michele Michelotto) wrote:
>You're comparing the top QIC format with the rather new DLT 
tecnology.  What is so special about DLT? I'll try to answer:

>1. Serpentine format means that there are several parallel tracks.
the head goes down the first track and comes back down the second one
etc.  If I need to access a file at the end of the "logical tape" and
the drive knows that it is at the beginning of the 52th track it goes
directly to the 52th track and start seeking on it. So the worst case
access time is close to the rewind time (about 100 sec) the average
access time is about (60 sec).

>2. the unit I tested was a 6 GB/cartridge (no compression) 700
kB/sec. the cartridge had 112 tracks but since the drives use two
heads, it could access track N and track N+54 together. So it looked
like a 54 track cartridge. Now it's very easy to put 4 (or 8) heads in
the drive and double (or multiply by 4) the transfer speed while
maintaing the backward compatibility (with 8 head you use only head #1
and #5 to emulate a 2 heads unit).

>3. DEC is selling to the OEM a DLT4000 unit with 20GB uncompressed
(40GB with compression). [may be available 9/94].  [Thinner, longer
tape plus somewhat higher density and slightly more efficient
packing/blocking]

DLT cost $5K US for 20GB drive, $10K for a 140GB stacker, $150K for a
3.2 TB robot.

Autochangers are made by DEC, Odetics (available through EMASS) and
Metrum (now MountainGate).

Piping tar into dd, with a bs=64k can increase your speed.

The drives have a tape mark directory that is used for a SPACE
command, but if you just SPACE 1 FILEMARK multiple times, efficiency
is poor (and is the fault of the software implementation as it should
"SPACE n FILEMARKS").

Submitted (approved?) ANSI standard, but that does not mean anyone
other that DEC is doing anything more than OEM'ing it.

------------------------------
Subject: [5.10.0.1] DLT4000 (From DEC)
From: Tape


Streaming tape drive
Quoted 40 GB includes 2:1 compression
Quoted 3.0 MB/sec includes 2:1 compression
Extended 5.25-inch form factor
SCSI-2 interface, either single-ended or differential, 
optional fast SCSI.
Compression is DLZ (Digital Lempel-Ziv)
"...a head life of 10,000 hours (compared to 2,000 hours for 
other tape products), a recommended average of 10,000 read/writes per 
cartridge, and an MTBF of 80,000 hours."
Search speed averages 68 seconds
Repositioning time 1.3 seconds
Hard error rate: 1 x 10**17 bits read
Undetected error rate: 1 x 10**30 bits read
Serial serpentine (128 tracks), variable block
bits/inch: 82,000
Tracks/inch: 256
Recording media: CompacTape (tm) IV
0.5 in x 1,700 ft x .3mi thick
Cartridge: 4.1 x 4.1 x 1 inch
shelf life: 10 years
Height: 3.235 in, width: 5.7 in, length: 9 in
Reliability: 80,000 MTBF
Media reliability: 500,000 passes in start/stop mode (or an 
average of 10,000 uses/cartridge)


------------------------------
Subject: [5.10.0.2] DLT2700 (from DEC)
From: Tape

random access, seven tape, 1 drive library
rack mountable 8-inch form factor
includes operator control pannel and LED indicators
400,000 mean mechanical cycles before failure
uses a 7 cartridge magazine.
Magazine "precheck": 75 seconds per magazine
Cartridge load (max): 28 seconds
Cartridge unload (max): 30 seconds
SCSI command set for robotic commands
Subsystem reliability: 30,000 power-on hours
Height: 10.4 in, width: 8.7 in, length: 27 in
weight: 65 lbs


------------------------------
Subject: [5.10.0.3] DLT2000 (from DEC)
From: Tape

CompacTape (tm) III
Capacity: 20 GB/cartridge
        (assumes 2:1 compression)
1.25 MB/sec.


        rdv@alumni.caltech.edu (Rodney D. Van Meter), vanepp@fraser.sfu.ca 
(Peter Van Epp), michelotto@mvxpd5.pd.infn.it, jeff@wsm.com, 
rrohbeck@ufhis.enet


------------------------------
Subject: [5.11] MountainGate (was Metrum) VHS {brief}
From: Tape

.MountainGate
.800-556-0222 or 702-851-9393 Phone
.702-851-5533 Fax
.in Europe: Mountaingate Data Systems Linda Radley in the UK
.+44-1256-464-767 (tel) +44-1256-597-48 (fax)
        Drive: RSP-2150, 2MB/sec sustained, 4MB/Sec burst
        ST-120 cartridge holds 14.5 GB
        ST-160 cartridge holds 18 GB
        See also robotics


Integrated with lots of SW packages, and drive prices have come way
down. I think you can now get them for less than $15K. Integrates
smoothly with robotics.

See http://www.nml.org/publications/NML_TR/lcmd/lcmdtoc.html for one
performance evaluation. Lots of good info there, but keep in mind that
the testing was conducted in 1992.

------------------------------
Subject: [5.12] VCR VHS
From: Tape

This appeared recently in the newsgroup, but I don't know anything
about it. This is the widget that takes data in one end and spits out
a video signal that you can pipe into your home VCR to use your it for
data storage. It's only $350, but, for those of us in the U.S. and
Japan, it doesn't work for NTSC VCRs.

From: simakov@glas.apc.org
Subject: VTS Tech Specifications for Users
Date: Mon Nov 28 09:24:29 PST 1994
X-Gateway: notes@igc.apc.org

Having received a lot of questions from different users of VTS 1020,
I'd like to answer them giving short specifications of this unit.
1. CAPACITY: 4 GB Compressed / 2 GB Uncompressed on one E-180 cassette.
             The amount of data grows according to the tape length.
2. SPEED:    100 KB/sec for PC/AT 286-16 Mhz
             200 KB/sec for PC/AT 386-33 Mhz etc.
3. SOFTWARE: The current version is for DOS. Windows support-DOS Window.
             Read/Write verification is provided.
4. VIDEO:    PAL/SECAM System, VHS Tape


------------------------------
Subject: [5.13] 19MM (D1 and D2) {Brief}
From: Tape

19mm is 3/4 inch helical scan tape.  Two varieties exist, D1 and D2.
Both originated from broadcast and/or data recorder applications,
where the data/signal was analog in nature.  They have been modified
for digital use, with error correcting capabilities added.

Data rates are in the 8-45 MByte/sec range, with storage capacities
in the 25-175 GB range in physically different size cartridges with
different length tapes, but all fitting into the same tape drive unit.

------------------------------
Subject: [5.14] ID-1
From: Tape

SHMO: I'd like to hear more of people's experiences using these
things.

Note: I would recommend you talk to people who've used these things
before buying one!

DataTape (Pasadena, CA, formerly a division of Kodak), I believe, has
an ID1 system available, with a HiPPI interface.

Sony makes several models of a D-1-based data drive; the format is
generally referred to as ID-1. It comes in different models, with
equivalent price tags, that run from 8 to 32 MB/sec. The original
machine had a VME interface that was extremely low-level ("any lower
and you'd have to turn the spindles by hand," someone said); now there
is a HiPPI interface available from a company called TriPlex. I
understand the HiPPI interface also adds another layer of ECC to
improve the otherwise abyssmal error rate (10^-10 becomes ???). Sony
is also supposed to be doing their own SCSI and HiPPI interfaces. I
don't know the status nor if they are compatible with tapes from the
TriPlex unit (I suspect not).

SONY has announced 3 interfaces so far. ;-)

DFC-1500 - SCSI interface

DFC-1700 - FW-SCSI-2 interface

DFC-1800 - 8 bit ECL interface that acts as a "scrollable" buffer.

All of the interfaces are fully buffered. They list for 40K-88K.

I have used the DFC-1700 for some time. From the standpoint of SCSI
functionality it is quite good. It obeys the MT commands and if you wish
you can either write to it in "raw" mode or DTF, which is SONY's version
of the DD-1 spec. On an HP I get 14+MB/Sec on the DTF side and near 16MB
in RAW mode.          

TriPlex makes an interface that does SCSI,FW-SWCI-2 and HIPPI. The price
varies. List for the HIPPI version is in the 130K range. The controller
is fully buffered.

Myriad Logic also builds boards (VME) and a controller. The controller
was demonstrated in Europe this year. It is a HIPPI attached controller
built out of their existing products. No good feel for the price, but I would
GUESS that it will be less than 90K. It is also fully buffered with 384MB
of VSB memory.

These are very expensive -- $100K+, but for people with the need,
they've got the speed.

        (stephens@access.digex.net, (John Stephens),
.rdv@alumni.caltech.edu (Rod Van Meter),
.sp@beta.lanl.gov (stephen w. poole) (8/95))

------------------------------
Subject: [5.15] D-2
From: Tape

Ampex (and Sony??) support D2.  Data rates are in the 15 MByte/sec
range.

The Ampex tape transport and head system were sold through E-Systems
(EMASS), who built the storage controller and sold it as the ER-90 and
coupled it with Odetics robots. Ampex now makes their own interface
for the unit, sold as the DST. They also make their own very fast
robotics.

The ER-90 is popular with the oil crowd. I don't know if the tapes are
interchangeable with the DST.

(rdv, 12/94)

DD-2  (19mm Data D-2 Format)

Ampex DST General:

3 cartridge (cassette) sizes - 25, 75, 165 gigabytes (uncompressed).
15 megabyte/sec. sustained (20 megabyte/sec. burst) transfer rate (per
drive).  Up to 800 megabyte/sec. search speed (per drive).  Smart DD-2
format includes partitioning and system zones to maximize storage
efficiency and speed data access.  3 layers of Reed-Solomon error
correction, with read-after-write verification and automatic rewrite
yield error rate of 1 in 10E15 bits read.  Drive(s) dual ported SCSI-2
(16 bit fast, differential).

DST 310 Tape Drive:

All 3 cartridge (cassette) sizes supported - 25, 75, 165 gigabytes
(uncompressed).  Rack mount or table top configurations.  Single unit
price: $120K.

Ampex Corporation
401 Broadway, M.S. 3-46
Redwood City, CA 94063-3199
Inquiry:  415-367-2982
Facsimile:  415-367-3850
Internet:  dst_mktg@ampex.com

(see also Ampex under autochangers -- they make their own for this
tape drive)

(pete_zakit@ampex.com, 94/12/23)

------------------------------
Subject: [5.16] StorageTek Helical {Brief}
From: Tape

Storage Tek has been working on a project called Redwood for a number
of years. The cartridge will be 3480 form-factor, to protect users'
investment in Storage Tek robotics. (rdv@alumni.caltech.edu (Rod Van
Meter)) Cartridges come in three lengths, with capacities of 10, 25
and 50 GB. That would put one of their 6,000-cartridge silos up to 300
TB.

The ESCON interface is in betatest; SCSI fast & wide due out soon.
(martin@viper.desy.de, 94/12/19)

Reportedly available now (95/5/15) with the SCSI interface; ESCON has
been delayed until end of the year.

Sustained transfer rate of 11.25 MB/s. Supports compression. List
price ~$100K.

Compatible with most of the STK robots. See http://www.stortek.com and
also autochangers in part 2 of this FAQ.

------------------------------
Subject: [5.17] Optical {new}
From: Tape

A company called Creo, from Canada I believe, makes a large tape drive
that uses ?1"? tape and gets a terabyte of data on a $10,000 reel.
The time to read the media (media granularity) is huge; at 3
MB/sec. it takes almost four days to read a tape!

EMASS have aquired the rights to manufacture the optical tape drive
from CREO. The drive ($350k) provides 1TB on line with data transfer
at 3MBytes/sec and an average seek time of just 30 seconds. Optical
tape media is supplied on 12.5" reels (capacity 1 TByte) by ICI
Imagedata in the UK. Expected archival life 30yrs.  Typical price of a
reel is $8,500. (updated 1995/9/18,
Duncan_Riddle@ici-imagedata.ccmail.compuserve.com)

------------------------------
Subject: [5.18] D-6 {brief}
From: Tape

From Toshiba & BTS, originally designed as a full-speed (~150MB/sec.)
digital HDTV VTR. A model with a HiPPI interface is supposed to be
available end of 94. The video version is priced at US$300K+. I
believe the tape transport and cartridge are the same as for D-2,
though the tape material is different.

(4/94, rdv)

------------------------------
Subject: [5.19] D-3 {brief}
From: Tape

From Martin Marrieta, mentioned here on the net recently. Very fast
(10.8 MB/sec.), ~$125K. General availability scheduled for 11/14/94
(9/20/94, garyblk@aera.com). I believe the cartridge is the same as
Betacam, so look for the broadcast autochanger companies here (rdv).

------------------------------
Subject: [5.20] INLINE RES-400 {Brief,New}
From: Tape

INLINE RES-400 is also a helical scan device, utilizing 1.2" Digital
Data Cartridges incorporating DTF (also brought from the broadcast undustry
with a lenghty history of success), capacitities of 12 and 42 Gigabytes
UNCOMPRESSED at a sustained transfer rate of 12.2 MB per second again
UNCOMPRESSED with and error rate of 10 to the -17!!  MTBF 200,000 HRS
(42 Gigabytes in less than 60 minutes W/O compression) Cost $55,000 per
transport with 16-bit SCSI differential interface.

800.465.4637 or 703.478.0800 main, 703.478.0966 fax

(John Tibbitts, john@inlinecorp.com, Oct. '95)

------------------------------
Subject: [6] Disk
From: Disk

------------------------------
Subject: [6.1] CAV, ZCAV and CLV
From: Disk

Many disks (hard, floppy and optical) run in CAV (Constant Angular
Velocity) mode. In this case, the disk spins at a constant rate, and
there are the same number of sectors per track on inner and outer
tracks. This means that the bits are farther apart on the outer
tracks, potentially wasting space. The transfer rate is
constant, as the number of bits/track is same and the time/track
doesn't vary.

CDs (and video laser disks, I believe) and early Macintosh floppies
run at Constant Linear Velocity (CLV). That is, the bits are all
roughly the same size, and the rotations per minute of the drive is
adjust as the head moves in and out. This gives the best areal
density of bits, at the sacrifice of seek speed, since every seek
requires an adjustment of the rotation speed. The transfer rate is
constant, as the size and spacing of bits is constant and the linear
velocity is constant.

The current rage is ZCAV, Zoned Constant Angular Velocity. Most modern
SCSI disks have this feature, and the newest MO drives do, as well.
There are a number of zones defined on the disk. The number of sectors
per track is different in each zone. Thus, the data is packed more
densely than normal CAV, but seek speed is not sacrificed. Another
effect of ZCAV is that the media transfer rate varies depending on the
head position, because the time/track is constant and the bits/track
vary; for example, the Seagate ST12450W Barracuda drive varies from 68
to 113 Mbits/sec, almost a factor of two different.

------------------------------
Subject: [6.2] Optical {Brief}
From: Disk
        See also Robotics section for library options.
        Slower than magnetic disks (in general)

------------------------------
Subject: [6.2.1] CD-ROM
From: Disk
Historically produced off-site at significant first-copy cost but 
small cost for high volumes.  Now on-site 'authoring' systems are available.  
Standard formats are available. Ads have been posted to the net
offering services for as little as US$60 to convert a tape to CD.

------------------------------
Subject: [6.2.1.1] Next-Generation CD-ROM {New}
From: Disk

There are two new standards in the works, digital video disk (DVD) and 
high-density compact disk (HDCD).  DVD is proposed by 8 major consumer 
electronics giants and will feature a double sided disk capable of 
storing 5GB of data per side.  HDCD backed by Philips and Sony will hold 
3.7GB data, with the potential to double them up to hold 7.4GB by
using a two-layer technology.

Now it is *Toshiba* that teamed with Time Warner on Digital Video Disc
(DVD). Sony and Philips have developed High Density Compact Disc
(HDCD).

The target transfer rate for both is MPEG-II, MP@ML (15Mbits/sec., I
think???), I believe, but may be lower (8 Mbps, perhaps???).

VHS/Beta wars all over again, along with issues such as backwards
compatability to existing CDs.

Anybody got an expected release date?

(John Wiest (john.wiest@24stex.com), gold@sri.com (MIchael Gold) and
others, 95/04/20)

------------------------------
Subject: [6.2.2] WORM {brief}
From: Disk
Write-Once-Read-Many
Standards are less firm between vendors.

------------------------------
Subject: [6.2.2.1] Sony {brief,new}
From: Disk

In the July '95 Wired, p. 56, and seen at SIGGRAPH '95:

new Sony WORM drive, looks like 12", maybe bigger. 15GB on a $400
platter. drive is $21K. 76-disk autochanger is $112.5K. ZCAV. transfer
rate 2.7 MB/sec. sustained.

call 1-800-222-7669.

(rdv,7/95)

------------------------------
Subject: [6.2.3] Erasable
From: Disk
Better standards than WORM

------------------------------
Subject: [6.2.3.1] Magneto-Optical Physics
From: Disk

Magneto-optical disks are plastic or glass disks coated with a
compound (often TbFeCo) that has special properties. The disk is read
by shining a low-intensity laser (originally infrared, but experiments
are being conducted all the way up to blue, I believe; the shorter the
wavelength the higher the possible density, all things being equal
(which they never are)) onto the media and examining the polarization
of the reflected light. To write, a higher-intensity laser is used to
heat the material up to its Curie point, where it becomes susceptible
to a magnetic field. When the media cools again, its state is
"frozen". The polarity of the reflected light during a read depends on
the polarity of the magnetic field under which the media was last
cooled. Once it has cooled it is no longer suceptible to magnetic
fields. Thus, it can be compared in a sense to paleomagnetism.

------------------------------
Subject: [6.2.3.2] Magneto-optical, 5.25-inch
From: Disk
Same number of sectors on each track whether or not track is near 
center or outer edge.  640 MB. Made by IBM, HP, Sony, Ricoh, others?

------------------------------
Subject: [6.2.3.3] Magneto-optical ZCAV, 5.25-inch
From: Disk

        Zoned constant angular velocity - more sectors on outer tracks.
        GB: 1
        ECMA standard 183 going through ISO Fast Track.
        ADSTAR demonstrated (6/93), San Jose, CA, 408/256-7895.


------------------------------
Subject: [6.2.3.4] HP Corsair {Brief, New}
From: Disk

1.3 GB on a double-sided cartridge.

See also under MO autochangers in part 2.

http://www.dmo.hp.com/storage/optical/main.html describes HP products.

------------------------------
Subject: [6.2.3.5] Maxoptix T4-1300
From: Disk

Does 1.3, 1.0 (read only) and 650 MB media. Max sustained read 2.0
MB/sec.

See also under MO autochangers in part 2 for contact info.

(rdv, 95/02/14)

------------------------------
Subject: [6.2.3.6] Asaca HSMO {Brief}
From: Disk

My company (Asaca) makes a 12.24 MB/sec. MO drive that uses custom
media and two four-beam heads in parallel to increase the transfer
speed. Expensive.

Call John Clemens in our L.A. office, (310)827-7144

------------------------------
Subject: [6.2.3.7] Other Multi-beam MO {None}
From: Disk

------------------------------
Subject: [6.2.3.8] 3.5-inch MO {Brief, New}
From: Disk

Generally looks like a slightly overweight floppy. All current ones
are single-sided.

First generation 3.5 MO was 128 MB on a cartridge.

Second generation devices (available now) are 230 MB.

Third generation (due out this year?) will be 650 MB.

If they bring the drive price down, could displace floppies as the
basic shirtpocket-transportable medium.

(rdv,95/1/20)

------------------------------
Subject: [6.2.3.8] Sony {Brief, New}
From: Disk

Sony now has 650 MB/side 3.5" MO with direct-overwrite capability (a
major step forward in MO). See
http://www.sony.co.jp/TechnoGarage/HS/index.html. It's ZCAV, and
nominally 1-2 MB/s. There's a reasonable discussion of MO physics and
technology here, too.  (march@media.mit.edu (Mark Holzbach), 10/95)

------------------------------
Subject: [6.2.3.9] Nikon 12-inch MO {Brief}
From: Disk

Holds 8 GB on a disk, with a transfer rate of ?>1MB/s.

------------------------------
Subject: [6.2.3.10] Sony 12-inch MO {Brief,New}
From: Disk

Sony also makes a 12" MO. 3.2 GB? (rdv,95/2/7)

------------------------------
Subject: [6.2.3.11] NEC 12-inch MO {Brief,New}
From: Disk

NEC also makes a 12" MO. (rdv,95/2/7)

------------------------------
Subject: [6.2.4] Electron-Trapping {None}
From: Disk

------------------------------
Subject: [6.2.5] Dual Function {Brief, New}
From: Disk

Capable of using both WORM and Erasable media. Some do the WORM in
firmware -- the media is really rewritable. Others do true WORM.

------------------------------
Subject: [6.2.5.1] Panasonic/Toray {Brief, New}
From: Disk

The Panasonic/Toray phase-change drive **READS** CD's but **WRITES**
phase-change discs which are not compatible with CD's and cost as much as
M-O media. (Mike Schuster, schuster@panix.com, 8/95)

------------------------------
Subject: [6.2.5.2] IBM {None}
From: Disk

------------------------------
Subject: [6.3] Magnetic
From: Disk

------------------------------
Subject: [6.3.1] 5.25-inch
From: Disk

------------------------------
Subject: [6.3.1.1] Seagate
From: Disk

Seagate's Elite 9 is 9GB -- reports here of backordering, others of
availability. Micropolis due out with an 8GB soon? (94/9/1)

The fastest (in sustained transfer rate) known 5.25" disk drive is the
ST12450W2HP 1.78GB Barracuda drive from Seagate. The Barracuda family
is large, so pay attention to the model number! It runs at 68-113
Mbits/sec., depending on head position (it's ZCAV). Assuming that data
rate is pre-format, and subtracting 20% for the format overhead, that
would be a sustained rate of 6.8-11 MB/sec. Of course, your mileage
WILL vary according to transfer size, locality, etc. (rdv,95/2/7)

------------------------------
Subject: [6.3.2] 3.5-inch
From: Disk

------------------------------
Subject: [6.3.2.1] IBM
From: Disk

The IBM DCMS-310800 Ultrastar2 is 10.8 GB (1GB=10^9) after format, and
its sustained rate is fast -- 8.4-14.2 MB/sec (presumably pre-format,
so subtract 20%). Only 5400 rpm with an 8.9 msec seek time, so
middle-of-the-pack on those numbers. Fast/wide SCSI-2 interface.
(rdv,95/02/14)

------------------------------
Subject: [6.3.3] Bernouli {None}
From: Disk

------------------------------
Subject: [6.3.4] Floptical {Brief}
From: Disk

I believe flopticals use an optical tracking mechanism to improve
ordinary magnetic head positioning and therefore density.

The Compaq/3M/Matsushita floptical floppies actually hold 120MB
formatted (according to the August 95 Byte), and can read and write
standard 1.44MB floppies and read 720KB floppies. (John Brock,
jbrock@panix.com, 8/95)

Supposed to be available Dec. '95?

------------------------------
Subject: [6.3.5] PC Removables {Brief, New}
From: Disk

SHMO -- I haven't followed this too closely. There's the Zip, Jazz,
and the new SyQuest.

------------------------------
Subject: [6.3.5.1] SyQuest EZ135 {Brief, New}
From: Disk

As of August 1995, $199.00 for the drive, $20.00 for each cartridge
(135 MB). Removable magnetic hard disk. SyQuest has been in the
business for years; the 5.25" removables were popular with Macs.

------------------------------
Subject: [6.3.5.2] Iomega Zip {Brief, New}
From: Disk

100 MB/disk, ~$200 for the drive, ~$15-20 for disks. SCSI or parallel
interface, 1.5 MB/sec. transfer rate? Don't know anything about the
technology, I assume it's simple magnetic hard drive. (rdv, 12/95)

------------------------------
Subject: [6.3.6] Mainframe {Brief}
From: Disk

Mainframe disks are sometimes referred to as SLEDs (Single Large
Expensive Disks). The term DASD (Direct Access Storage Device) usually
refers to a mainframe disk, but is occassionally applied to any hard
disk.

The WWW FAQ contains some information about mainframe CKD disks.

------------------------------
Subject: [6.4] Other
From: Disk

------------------------------
Subject: [6.4.1] Holographic Storage Products {Brief}
From: Disk

Tamarack Storage Devices, Inc, a spin-off from Microelectronics and 
Computer Technology Corporation (MCC), is developing with Projectavision Inc. 
to produce a product with ten times greater storage densities than magnetic 
and 10 to 1000 times faster than floppies, tapes, and CD-ROMS.  First 
products expected first quarter 1994.
        (Ref: MCC Collagorations Newsletter, Volume 3, No. 1; Spring 1993)

        (stith@fnal.gov)

Also another report of experiments at Stanford recently (8/94).

Scientific American in the Nov. 95 issue reportedly has an article
about holographic storage, but I haven't tracked it down yet.

------------------------------
Subject: [7] RAID {Full}
From: RAID {Full}

The primary functions of a disk array is to increase data 
availability, to increase total storage capacity, and to privide performance 
flexibility by selectively spreading data over multiple spindles.

Data Protection - As the number of disks on a system increases, the 
likelyhood of one failing increases.  Thus, a disk array should be immune 
from a single disk drive crash.  Disk mirroring (keeping an exact copy of a 
one disk on another) is the simplest, but requires twice the disk capacity 
(and associated cost).  Encoding schemes can be used to reduce the redundancy 
required to lower ratios.

Storage Capacity is increased by placing many smaller form factor 
(5.25 and 3.5-inch) drives onto an intellegent controller which makes all the 
drives appear as one drive to the computer system.

Performance can be increased by spreading data over spindles and 
performing operations in parallel which allows multiple drives to be working 
on a single transfer request.

------------------------------
Subject: [7.1] RAID Levels
From: RAID {Full}

The original taxonomy of RAID levels was published in the SIGMOD paper
by Garth Gibson and Randy Katz in 1988 (see below). The taxonomy
roughly classifies RAID architectures according to the layout of data
and parity information on disks. It is NOT gospel and does NOT cover
every possible architecture (it has been pointed out here that that
would require an N-tuple showing data block addressing, number and
types of parity and ECC information, etc.), but when used properly
provides a vocabulary and establishes a framework for discussion.


        Raid Level 0 - Striping - Data is segmented and split onto multiple 
spindles.
                Short Reads - Easily handles multiple simultaneous reads
                Long Reads - Single operation can be split and processed in 
parallel
                Short Writes - Easily handles multiple simultaneous reads
                Long Writes - Single operation can be split and processed in 
parallel
                Redundancy - None
                Cost - Good (no extra hardware)
        Raid Level 1 - Mirroring - Duplicate data is kept on multiple 
splindles
                Short Reads - Faster (shorter latency) since
..resolution can be from any of multiple disks
                Long Reads - Faster since resolution can be from any of 
..multiple disks (*)
                Short Writes - Slower since need to write to multiple disks
                Long Writes - Slower since need to write to multiple disks
                Redundancy - Excellent
                Cost - Expensive - at least double the spindle cost
        Raid Level 3 - Data protection disk - mathematical ECC type code 
calculated from multiple spindles and stored on another spindle.
                Short Reads - Normal speed (i.e. 1x per-spindle rate)
                Long Reads - Normal speed
                Short Write - Slower due to re-calculating of ECC code 
(including reading from other spindles and the ECC write)
                Long Write - slightly slower due to ECC writes, but less 
reading required than in short writes (**)
                Redundancy - Excellent
                Cost - only slighly more than no redundancy options
.Raid Level 4??? similar to 3, with block striping instead of byte.
        Raid Level 5 - Striping plus data protection - stripe data across 
multiple spindles (as in RAID Level 0) and have data protection calculations 
(as in RAID level 3) but don't put all the calculated figures onto one 
spindle, but spread it out.
                Short Reads - Normal
                Long Reads - Faster due to parallelism
                Short Write - Slower due to ECC calculation (including 
reading and writing)
                Long Write - slighly slower due to ECC writes (**)
                Redundancy - Excellent
                Cost - only slignly more than no reduncancy options

(* should be the same speed as a single spindle)
(** -- should be faster than a single spindle due to parallelism on
write? somebody help me out --rdv)

        Benefits of RAID:
                High data availability (ie, if a single spindle crashes, no 
data is lost)
                Increased disk connectivity per system  - since multiple 
spindles appear as one spindle to the computer system.
                Large capacity storage in a small footprint - 
                Flexibility through intelligent array controllers
                Performance enhancements in some circumstances.


Streamed or Streamified RAID??? (SHMO)

------------------------------
Subject: [7.2] RAID-6
From: RAID {Full}

A two-dimensional disk array parity scheme was described by Randy Katz,
Garth Gibson, and David Patterson (all then with UC Berkeley - Gibson is
now a professor at Carnegie Mellon University) at the 1989 IEEE Compcon
conference. This method had one parity calculated along the disk strings
and another calculated across them. This would increase the
mean-time-to-data-loss by more than 10,000 fold. I am not aware of any
implementations of this configuration.

Storage Technology Corp (STK - Louisville, Colorado) has described a
somewhat similar scheme for their long-delayed Iceberg disk array. This
would have a regular, orthogonal RAID 5 parity across drives along with a
Reed-Solomon encoding on another drive. This is sometimes referred to as
RAID 6 or RAID 5+. STK claims their design will allow failure of ANY TWO
drives - which is beyond the survival capabilities of standard RAID 5.

A RAID 5 which is 'deep' can survive failures in more than one drive so
long as it doesn't lose more than one drive per rank:


        HBA1    HBA2    HBA3    HBA4    HBA5    HBA6    HBA7    HBA8
         |       |       |       |       |       |       |       |
Rank1   Disk1   Disk2   Disk3   Disk4   Disk5   Disk6   Disk7   Disk8
         |       |       |       |       |       |       |       |
Rank2   Disk9   Disk10  Disk11  Disk12  Disk13  Disk14  Disk15  Disk16

        . . . . .   - etc.

Rank4                           .       .       .       .       Disk32


If the above is a RAID 5 then losing drives 5 & 6 will destroy data. If it
is a RAID 6 then it will not. Losing drives 3 and 12 will not disable a
RAID 5 nor a RAID 6.

But RAID 6 will cost more and may have slower performance for small random
writes from having to update more parity data. I think there are clearly
ways to mitigate the parity update perfomance for RAID 6 as well as RAID 5.

-- 
          Dick Wilmot
          Editor, Independent RAID Report
          (510) 938-7425



------------------------------
Subject: [7.3] John O'Brien and RAID-7
From: RAID {Full}

RAID-7 is a marketting term created by Storage Computer, Inc. for what
others here have described as RAID-4 with a write cache. John O'Brien
(RAID7@world.std.com), (their marketting manager?) frequently posts
here. 

His claims of ~10x improvement on I/O rates for VAXes have been shown
to be poorly measured; the change in systems was not simply a
RAID-for-modern-disk swap, but included increasing the CPU power by a
factor of three and eliminating the HSC and old disk technology. He
has also made difficult-to-substantiate claims about the growth and
market success of his company relative to competitors. Thus, wise
advice would be to take everything Mr. O'Brien says with a grain of
salt (not bad advice for dealing with anyone, but especially true for
dealing with vendors).

The debate also appears here frequently as to whether or not you
really WANT your RAID array doing write cacheing; Unix file systems
may depend on specific ordering of writes and otherwise make
assumptions that could leave you in trouble with power or disk
failures. If write ordering is preserved, the danger is somewhat
mitigated.

That said, some posters here are pleased with their RAID7 arrays, and
although comp.arch.storage opinion runs prevalently against Mr.
O'Brien himself (and lately his pal Michael Willett who interestingly
is quoted here from before he worked for Storage Computer), the
possibility exists that the product is worthwhile.

------------------------------
Subject: [7.4] RAID Papers
From: RAID {Full}

(Berkeley FTP pointers updated, 95/5/11)

A nice collection of RAID papers was published in the Fall, 1991 issue of
_CMG Transactions_. A few more appeared in the December, 1992 _CMG
Proceedings_ and there are 3 RAID papers in the 1993 International
Symposium on Computer Architecture (Published as _Computer Architecture
News 21_, #2, May, 1993 by ACM SIGARCH.

(dwilmot@crl.com, Dick Wilmot, Editor, Independent RAID Report)

Try contacting the RAID project at the University of California, Berkeley. 
In the proceedings of the recent IEEE Mass Storage Symposium, Ann Drapeau
and Randy Katz have a paper describing the reults of some investigations
into the use of tape arrays.  I think you can find RAID papers, perhaps
this one, on anon ftp at ftp.cs.berkeley.edu.  Have no address for Ann
Drapeau, but Randy Katz is randy@cs.berkeley.edu.

Some of the RAID papers are available via anon ftp from
ftp.cs.berkeley.edu:pub/raid/papers

Ann Drapeau's email address is alc@cs.berkeley.edu.

(dm_devaney@pnl.gov, Mike DeVaney)
(eklee@cs.berkeley.edu, Edward K. Lee)

>>I am looking for papers or technical papers on RAID...

You could get that lengthy RAID taxonomy research report from Storage
Computer as mentioned recently on these news groups, by Emailing them at
RAID7@World.std.com    Alternatively, their phone number is 603 880 3005.
I do not know if their RAID research report is copyrighted or not.

I believe their executive in charge of RAID activities in Hong Kong would be
John Taylor, the former Wang national accounts director.  They also put
on technical raid seminars which might be of interest to your PhD students,
concentrating on performance enhancements over RAID 3/4/5 (somewhat less than
an order of magnitude, but I have not reviewed their benchmark data.)  The
RAID theory discussed is rather interesting.

(MICHAEL.WILLETT@OFFICE.WANG.COM, Michael Willett)
---------
>> I am looking for papers or technical papers on RAID or other multiple disks
>> storage systems.  Could somebody give me pointers for them?

Here are some papers that I either have read or am looking for:
 
I don't have copies of this group:

Dishon, Yitzhak; Lui, T.S.; Disk Dual Copy Methods and Their Performance;
  FTCS-18: Eighteenth International Symposium on Fault-Tolerant Computing,
  Digest of Papers p 314-318

Gray, J.N. et. al., Parity Striping of Disk Arrays: Low Cost Reliable 
  Storage With Acceptable Throughput, 16th International Conference on
  VLDB (Austrailia, August 1990)

Katz, R.H.; Patterson, D.A.; Gibson, G.A.; Disk System Architectures for
  High Performance Computing; Proc. IEEE v 78 n 2 Feb 1990

Muntz, Richard R.; Lui, John C.S.; Proformance Analysis of Disk Arrays
  Under Failure; Proceedings of the 16th International Conference on Very
  Large Data Bases (VLDB); Dennis Mcleod, Ron Sacks-Davis, Hans Schek
  (Eds.), Morgan Kaufmann Publishers, Aug 1990 pp 162-173

Ng, Spencer; Some Design Issues of Disk Arrays; Compcon '89: Thirty-Fourth
  IEEE Computer Society Internationsl Conference p 137-142
  DISK ARRAYS, STRIPING, SPINDLE SYCHRONIZATION

Ng, Spencer W.; Improving Disk Performance via Latency Reduction; IEEE
  Transactions on Computers v 40 1 Jan 1991 p22-30
  LATENCY REDUCTION, ROTATION LATENCY, DISK PERFORMANCE

Reddy, A.L. Narasimha; Banerjee, Prithviraj; Performance Evalutaion of 
  Multiple-Disk I/O Systems; Proceedings of the 1989 International
  Conference on Parallel Processing p 315-318

Here are some good papers on disk arrays with emphasis on RAID:

Chen, Peter M.; Gibson, Garth A.; Katz, Randy H.; Patterson, David A.;
  Evaluation of Redundant Arrays of Disks Using an Amdahl 5890; 1990 ACM
  SIGMETRICS Conference on Measurement & Modeling of Computer Systems p 74-85

Chen, Peter M.; Patterson, David A.; Maximizing Performance in a Striped
  Disk Array; Proceedings of the 17th IEEE Annual International Symposium on
  Computer Architecture p 322-331

Chen, Shenze; Don Towsley; Performance of a Mirrored Disk in a Real-Time
  Transaction System; 1991 ACM SIGMETRICS Conference on Measurement &
  Modeling of Computer Systems p 198-207

Chervenak, Ann L.; Katz, Randy H.; Performance of a Disk Array Prototype;
  ACM SIGMETRICS 1991 Conference Proceedings p 188-197

Menon, J.; Mattson, R.L. and Spencer, N.; Distributed Sparing for Improved
  Performance of Disk Arrays; IBM Research Report RJ 7943 (Jan. 1991)

Patterson, David A.; Chen, Peter; Gibson, Garth; Katz, Randy H.;
  Introduction of Redundant Arrays of Inexpensive Disks (RAID); Compcon 1989:
  Thirty-Fourth IEEE Computer Society International Conference 
  p 112-117

Schulze, Martin; Gibson, Garth; Katz, Randy; Patterson, David A.; How Reliable
  is a RAID; Compcon '89: Thirty-Fourth IEEE Computer Society International
  Conference p 118-123


(danj@hub.parallan.com, Dan Jones)
--------
>>I am looking for papers or technical papers on RAID...

A good set of the Berkeley papers are available via anonymous FTP.
If I remember, the machine was ftp.cs.berkeley.edu.  Also, an archie
search on "RAID" would probably turn up a nice on-line collection of
information.  (sorry, not at an Internet site to check this right now...)

(buck@siswat.hou.tx.us , Lester Buck)

Further Information:
    %A Garth Gibson
    %A Randy H. Katz
    %T A case for redundant arrays of inexpensive disks (RAID)
    %C Proc. SIGMOD.
    %c Chicago, Illinois
    %D 1--3 June 1988
    %P 109 116
    %k RAID, disk striping, reliability, availability, performance
    %k disk arrays, SCSI, hardware failures, MTTR, MTBF
    %k secondary storage
    %L Jacobson has a copy
    %x Increasing the performance of CPUs and memories will be
    %x squandered if not matched by a similar performance increase in
    %x I/O.  While the capacity of Single Large Expensive Disks (SLED)
    %x has grown rapidly, the performance improvement of SLED has been
    %x modest.  Redundant Arrays of Inexpensive Disks (RAID), based
    %x on the magnetic disk technology developed for personal
    %x computers, offers an attractive alternative to SLED, promising
    %x improvements of an order of magnitude in performance,
    %x reliability, power consumption, and scalability.  This paper
    %x introduces five levels of RAIDs, giving their relataive
    %x cost/performance, and compares RAID to an IBM 3380 and a
    %x Fujitsu Super Eagle.

(tage@cs.utwente.nl)


------------------------------
Subject: [7.5] R-Squared {Brief}
From: RAID {Full}
Vangard Disk Array for DEC, Sun, HP, IBM RS/6000, SGI and others

Address: 11211 E Arapahoe Rd., Suite 200, Englewood, CO 80112

Phone: 303/799-9292, Fax: 303/799-9297

------------------------------
Subject: [7.6] Sun {Brief}
From: RAID {Full}

Sun Microsystems has a new Fibre Channel array that does RAID 0, 1,
and 5. See http://WWW.Sun.Com under the products descriptions.
(rdv,94/8/8)

------------------------------
Subject: [7.7] the RAIDbook {Brief,New}
From: RAID {Full}

The RAIDBook, a 100+ page tutorial on RAID technology and the
RAID Advisory Board, is available from Technology Forums, LTD,
of 6931 Glenview Lane, Lino Lakes, MN 55014-1296.

Contact Joe Molina, President of Technology Forums at 612-784-2379 or
612-784-0206.

I've read it, it's decent but a little repetitive. Defines many
low-level terms of interest only to those who need to know the
internals. (rdv,95/2/7)

------------------------------
Subject: [7.8] Software Striping {Brief}
From: RAID {Full}

Silicon Graphics provides software striping of SCSI disks; thus your
host can effectively act as a RAID controller, providing flexibility
and probably reduced price, possibly with a performance penalty in the
form of increased CPU overhead. However, it probably means that it can
spread the I/O load over multiple I/O controllers.

(similar features in other systems? SHMO --rdv)

RAID0 is in late beta under Linux. (evesg@etlcom3.etl.go.jp (Gjoen
Stein), 95/10/6)

sdsadmin on the HP 7xx line does raid 0 striping and works well.
this is also apparently possible on the 8xx machines using LVM.
sdsadmin is due to disappear with hpux 10, replaced by LVM.

I believe the Advanced FS on Alphas can also do raid 0.

(mark hahn, hahn@neurocog.lrdc.pitt.edu, 94/11/17)

ATTO Technology.has ExpressStripe, which does software striping on a
Mac.

Cyranix.http://www.cyranex.com makes EZRAID PRO (RAID 0,1,4,5) for
OS/2.  Voice: +1 613 738 3864 Fax: +1 613 738 3871

------------------------------
Subject: [7.9] RAID Vendors {New}
From: RAID {Full}

RAID vendors come and go quickly, OEM each other's equipment, change
names, and other activities that seem aimed at simply obscuring the
market. No list like this could be complete and up to date for long;
I'll gladly take updates.

The November '94 issue of _Advanced Imaging_ has a big article on
storage, primarily RAID arrays, with a pretty comprehensive list. This
table is distilled from that. Most of the info appears to be from the
vendors themselves. Almost all of these are fast/wide SCSI; a few are
Fibre Channel, NuBus, PCI or HiPPI (usually with IPI-3 command set). 
Most of these vendors have more than one model, only a few are listed
here. (rdv,95/1/18)

Most of these have some web presence; a Lycos search would turn up
their sites.

PC = Personal Computer (IBM compatible)
MC = Macintosh
PS = PC Server (Netware, NT et al)
UX = Unix (generic)
PU = Personal Unix
WU = Workstation Unix & workstation servers
MF = mainframe
MI = minicomputer (AS/400)
SU = Supercomputer

FC = Fibre Channel interface (usually SCSI command protocol)

Maker...Model..RAID Levels.Uses
-------------------------------------------------------------------------
AC Technology..Concorde.0,3,5..WU
ADJFILE Systems..Cougar, Tiger.0,1,3,5..??
AT&T Global Information.Series 3.??..WU,PS,PC
  Systems -- NCR
BusLogic..DA-x988..0,1,3,5..PC,PU,PS (PCI)
Canary Communications.IDA3500..0,1,3,4,5.??
Ciprico...6800 Real-Time.??..??
...RAID Array
DEC...StorageWorks.0,1,5..??
...RAID Array 210
Distributed Processing.SmartRAID.0,1,5..PC,PU,PS
  Technology
DynaTek Automation.AddARRAY.0,1..??
  Systems
Fujitsu Comp. Prod..DynaRAID.??..??
  America
FWB, Inc...SledgHammer*FT.5..MC
IBM Storage Systems.7137 Disk Array.0,5..WU
Legacy Storage Systems.SmartArray.??..PC (PCI)
Maximum Strategy.Gen5 Storage.0,1,3,5..SU (HiPPI,FC)
...Server
Mega Driver Systems.MR & MK Series.0,3,5..PC,PU,PS,MC,WU
MicroNet Technology.RAIDbank Plus.0,1,5..PC,PS,MC,PU?
Micropolis..RAIDION,GANDIVA.??..PC,MC,PS,PU,WU
Microtech Int'l..XLerator.0,1..MC
Mylex...DAC960S..0,1,5,6?,7?.??
Procom Technology.LANForce-5.0,1,3,5..MC,??
Raidtec...FlexArray IX.0,1,3,5..??
Recognition Concepts.RDR series.??..??
Storage Computer.RAID 7..7?(4?)..??
Storage Concepts.Concept 910.??..??
Storage Tek..Iceberg..??..MF
XL/Datacomp..9638..5..MI,WU

------------------------------
Subject: [8] Solid State Disk (SSD) Frequently Asked Questions (FAQ)
From: Solid State Disk (SSD) Frequently Asked Questions (FAQ)

*Note: This section is a slightly trimmed and editted version of the
SSD FAQ from Robert at DES (rdavid@ccmgate.mti.com) which I think
he also posts to c.d.sybase. I would take the "up to 1000 times
faster" claim with a grain of salt, though the general info is good.
--rdv, 94/9/15

1) Q. What are solid-state disk emulators?

A. Simply put, solid-state disk emulators are Dynamic Random Access
Memory (DRAM)-based storage devices that appear to the host exactly as a
magnetic rotating disk.  DRAM chips, which are ultra-fast devices
that store data while the system is on, increase data access, thereby
eliminating I/O bottlenecks that constrain overall system performance.

Solid-state disk emulators can be either volatile or non-volatile,
meaning that they are able to retain data when the system is turned off.
DRAM alone is volatile.  Solid-state disk emulators that are designed with an
integrated backup system are non-volatile storage devices; if a power
outage occurs, the user's data is protected by the backup system and will
not be lost. Solid-state disk emulators are volatile when methods for
backing up data are absent.  A power failure will cause data to be
lost on a volatile solid-state disk.

2) Q. How do solid-state disk emulators work?

A. Solid-state disk emulators plug into a computer's I/O controller.
Typical client/server systems use the ANSI-standard SCSI interface on
its I/O controller.  It is plug-and-play because it emulates a
rotating disk.  No special drivers or operating system patches are
required to make it work.  In addition, because there are no moving
parts, seek and rotational latency times are zero, which aids
solid-state disk emulators in performing up to 1000 times faster than
magnetic rotating disk drives.

3) Q. What applications are well-suited for Solid State Disk?

A. In general terms:  1) transaction processing, 2) batch processing, and
3) query or decision support analysis. Many types of application
software can take advantage of the super-fast access times SSD offers.

4) Q. How reliable are Solid State Disks?

A. Based on real world user data from a large SSD site, the actual power
on hours mean-time between failure is greater than 1,000,000 hours. Since
this site has yet to have a failure, the number is likely to go up.

------------------------------
Subject: [9] Other Devices
From: Other Devices

------------------------------
Subject: [9.1] SyQuest Removable Cartridge hard Drives
From: Other Devices
        form factor: 2.5", 42MB
        form factor: 3.5", 105MB, 14msec ave seek, 3600 RPM, ave sustained 
transfer rate: 1.3MB/Sec, available in IDE and SCSI versions. Syquest 
Technology, Inc., 47071 Bayside Parkway, Frement CA 93438, Phone: 800/245-
2278.

------------------------------
Subject: [9.2] Kalok removable cartridge hard drives
From: Other Devices
        3.5-inch form factor, 250 MB
        Phone: 408/747-1315

------------------------------
Subject: [10] RAIT (Redundant Arrays of Inexpensive Tape)
From: RAIT (Redundant Arrays of Inexpensive Tape)

(6/93)
There are at least two tape array products on the market:

Data General is selling the CLARiiON Tape Array Subsystem comprising
between five and seven 4mm DAT tape drives. Data can be recorded in
RAID-like striping redundancy, mirrored, or in conventional DAT layout.
This unit can provide up to 30GB of unattended contiguous storage. The
tape drives can record at sustained rates of 183 - 732 KB/second each but
customers should expect sustained backup at around 1 megabyte/second of
compressed data after accounting for host overheads. Data General is
working on a seven tape caddie to hold tape sets together. It is essential
that tapes in a RAID group not be separated.

NCR announced a tape array software product for NCR uniprocessors and
System 3450, System 3550 and StarServer Systems running UNIX V R4.2.01.
This tape array software yields faster and more reliable backup of large
database and file servers than with any single tape drive available today
but uses customers existing tape devices. It writes simultaneously to
multiple drives and can use array techniques to recover from loss or
failure in any single tape.

The motivations for tape arrays seems to parallel those for disk arrays:
        -       higher bandwidths
        -       increased reliability

(dwilmot@crl.com, Dick Wilmot, Editor, Independent RAID Report)

(6/93)
Pick up any DEC related trade rags and you can find an ad for an 8mm tape
array.  The ad I just found is by Contemporary Cybernetics and uses two
five GB 8mm drives with compression - they CLAIM to be able to get 50 GB os
storage total - but how many customers have 50 GB worth of 5:1 compressible
data?

Anyway - the ad doesn't mention RAID, but they support RAIDish (!) features
such as striping and mirroring.  It also supports offline tape-to-tape
copy and will automatically cascade onto the second tape when the first
one fills (useful for utilities that can't deal with multi-drive/multi-
volume).

I SEEM to remember someone having something like this with more drives, but
of course I couldn't locate the ad.

I would be really interested in seeing something like this for 3480 since the
transfer rate is already quite high...

(tbodoh@resdgs1.er.usgs.gov, Tom Bodoh)

At the Monterey IEEE Mass Storage Conference in April '93, Ann Drapeau
from Randy Katz's group presented a paper on striped tape.

The National Storage Lab High Performance Storage System reportedly
supports striping of removable media in the system software.

(rdv,95/1/13)

Something that came through the newsgroup recently (95/2/5):


Tape Arrays
High Performance tape drive units for large networks and minis.
Fast: up to 4Megabyes/second
High Capacity: from 24Gb on 4mm DATS to 60GB on DLTs; with 
autoloaders,up to 616GB
Flexibility: Stripe data across 4 drives, mirror data, 
stripe 2/mirror 2 - double your speed while creating an off-site
storage copy; off-line copy; pass-thru mode, etc.  
Transparent to your backup software - no changes or retraining
Compatible with all major OSs; including Novell, WindowsNT,
Unix, Sun, HP, Silicon Graphics, VMS, etc.
For More information:
William Wirth
Travlnmn@ix.net.com


Just spotted this in a PC rag. Andataco can stripe, mirror or RAID
DLT, 8mm or 4mm. Check out http://www.andataco.com or call
800-334-9191 or +1-619-453-9191. Or email inquire@andataco.com. 
(Andataco is an integrator for numerous storage products including
RAID arrays.)

------------------------------
Subject: [11] RAOT (Redundant Arrays of Other Things :-)
From: RAOT (Redundant Arrays of Other Things :-)

Pinnacle Micro has been advertising what they call the Orray --
essentially RAID done with removable magneto-optical disk drives. They
claim sustained transfer rates up to 8 MB/sec., which seems
implausible given that it's only four drives, and even HP MO drives
are below 2 MB/sec. sustained. Apparently no redundancy in the system
for that configuration (so I guess my ROAT designator is a misnomer),
but it should be generally less necessary for MO than magnetic disk
(drive failures normally don't result in the destruction of data or
media).

..--rdv, 94/7/20
