SATA drive on SCSI bus

PC Hardware
Dicussion of PC hardware and peripherals

FAQ

Memberlist

Profile

SATA drive on SCSI bus Goto page Previous 1, 2, 3 Next

PC Hardware Forum Index -> SCSI

Author

Message

Winey
Guest

Posted: Mon Sep 27, 2004 5:38 am Post subject: Re: SATA drive on SCSI bus-more questions about SAS

On Thu, 23 Sep 2004 23:33:40 -0600, "Phil Barila" <PBarila@Barila.com>
wrote:

Quote:

"Winey" <NOSPAMME@no-one-here.com> wrote in message
news:8ls4l0p4i7b313ka1ruf3mpoiph6s4num4@4ax.com...

[snip]

A while ago I read a posting about SATA and SCSI that said, AFAIR,
that SAS would also allow attachment of SATA drives. Somehow. What
do you think?

SAS does indeed allow attachment of SATA devices. The converse isn't true,
you can't attach a SAS drive to a SATA HBA.

I'm curious about this. Is there a URL that goes through these
issues, and explains what the benefits are of doing SATA on a SAS-host
adapter? Can SATA and S-SCSI devices be mixed on the same HA?

Also, picking up on your other posting about the mechanical
differences between SCSI and ATA, if I can attach SATA to a SAS host
adapter, why not implement all the SCSI commands, with an autosense
for the type of host adapter?

NCQ seems to be an important benefit. When you sift through all the
various religious arguments about SCSI vs. ATA for personal use
systems, there appear to be home systems, I think there are two
themes:

more devices on SCSI host adapter than ATA IRQ
better performance, which I think can be attributed to higher
performance mechanics of SCSI drives PLUS NCQ.
SCSI drives cost much more, up to 3X compared to 'equivalent' ATA
drives

So if SATA drives have NCQ, why not implement FORMAT UNIT and other
SCSI-only commands?

That would seem to be a way to address SCSI economics, while giving
people more choices.

(and this says nothing, by the way, about RAID).

Ron Reaugh
Guest

Posted: Mon Sep 27, 2004 6:02 am Post subject: Re: SATA drive on SCSI bus-more questions about SAS

"Winey" <NOSPAMME@no-one-here.com> wrote in message
news:garel05qcger2ul6ii6l2u6u5r0omd3jn6@4ax.com...

Quote:

On Thu, 23 Sep 2004 23:33:40 -0600, "Phil Barila" <PBarila@Barila.com
wrote:

"Winey" <NOSPAMME@no-one-here.com> wrote in message
news:8ls4l0p4i7b313ka1ruf3mpoiph6s4num4@4ax.com...

[snip]

A while ago I read a posting about SATA and SCSI that said, AFAIR,
that SAS would also allow attachment of SATA drives. Somehow. What
do you think?

SAS does indeed allow attachment of SATA devices. The converse isn't
true,
you can't attach a SAS drive to a SATA HBA.

I'm curious about this. Is there a URL that goes through these
issues, and explains what the benefits are of doing SATA on a SAS-host
adapter? Can SATA and S-SCSI devices be mixed on the same HA?

Also, picking up on your other posting about the mechanical
differences between SCSI and ATA, if I can attach SATA to a SAS host
adapter, why not implement all the SCSI commands, with an autosense
for the type of host adapter?

NCQ seems to be an important benefit. When you sift through all the
various religious arguments about SCSI vs. ATA for personal use
systems, there appear to be home systems, I think there are two
themes:

more devices on SCSI host adapter than ATA IRQ
better performance, which I think can be attributed to higher
performance mechanics of SCSI drives PLUS NCQ.
SCSI drives cost much more, up to 3X compared to 'equivalent' ATA
drives

So if SATA drives have NCQ, why not implement FORMAT UNIT and other
SCSI-only commands?

The primary HD mfgs make both SCSI and ATA HDs. They have to protect that
3x SCSI margin.

Quote:

That would seem to be a way to address SCSI economics, while giving
people more choices.

(and this says nothing, by the way, about RAID).

Phil Barila
Guest

Posted: Tue Sep 28, 2004 8:13 am Post subject: Re: SATA drive on SCSI bus-more questions about SAS

"Winey" <NOSPAMME@no-one-here.com> wrote in message
news:garel05qcger2ul6ii6l2u6u5r0omd3jn6@4ax.com...

Quote:

Absolutely. It's been demonstrated several times since April. I don't know
when it's commercially available. Google for:
SAS SATA demo.

Here's some shameless self-promotion:

http://www.seagate.com/cda/newsinfo/newsroom/releases/article/0,1121,2074,00.html

Quote:

SATA attachment to SAS puts all the burden on the SAS HA. The SATA drive
doesn't even know it's attached to a SAS HA.

Quote:

SATA NCQ isn't SCSI tagged command queuing. So adding FORMAT UNIT doesn't
produce a SCSI drive. FORMAT UNIT rewrites the sectors on the disk, even
resizing them. There is support for variable sector sizes in ATA7, but the
drives are only just getting there.

Quote:

That would seem to be a way to address SCSI economics, while giving
people more choices.

Enterprise drives often have one CPU dedicated to the servo, and another one
for everything else. ATA drives usually have one CPU for everything.
That's just one example of construction differences that contribute to the
cost delta.

Phil
--
Philip D. Barila Windows DDK MVP
Seagate Technology LLC
(720) 684-1842
As if I need to say it: Not speaking for Seagate.
E-mail address is pointed at a domain squatter. Use reply-to instead.

Ron Reaugh
Guest

Posted: Tue Sep 28, 2004 8:51 am Post subject: Re: SATA drive on SCSI bus-more questions about SAS

"Phil Barila" <PBarila@Barila.com> wrote in message news:sO-

Quote:

Enterprise drives often have one CPU dedicated to the servo,

Which current Seagate models have two CPUs onboard?

Quote:

and another one
for everything else. ATA drives usually have one CPU for everything.
That's just one example of construction differences that contribute to the
cost delta.

Phil
--
Philip D. Barila Windows DDK MVP
Seagate Technology LLC
(720) 684-1842
As if I need to say it: Not speaking for Seagate.
E-mail address is pointed at a domain squatter. Use reply-to instead.

Winey
Guest

Posted: Tue Sep 28, 2004 10:37 am Post subject: Re: SATA drive on SCSI bus-more questions about SAS

On Mon, 27 Sep 2004 22:13:23 -0600, "Phil Barila" <PBarila@Barila.com>
wrote:

Quote:

I'm curious about this. Is there a URL that goes through these
issues, and explains what the benefits are of doing SATA on a SAS-host
adapter? Can SATA and S-SCSI devices be mixed on the same HA?

Absolutely. It's been demonstrated several times since April. I don't know
when it's commercially available. Google for:
SAS SATA demo.

Here's some shameless self-promotion:

http://www.seagate.com/cda/newsinfo/newsroom/releases/article/0,1121,2074,00.html

No more shameless than any other hi-tech company bragging about its
latest "firsts." Everyone in the biz does this. (I also work in
hi-tech.)

Quote:

Also, picking up on your other posting about the mechanical
differences between SCSI and ATA, if I can attach SATA to a SAS host
adapter, why not implement all the SCSI commands, with an autosense
for the type of host adapter?

SATA attachment to SAS puts all the burden on the SAS HA. The SATA drive
doesn't even know it's attached to a SAS HA.

Oh. Does the SAS HA assign a SCSI ID to the SATA drive? LUN even?

Quote:

NCQ seems to be an important benefit. When you sift through all the
various religious arguments about SCSI vs. ATA for personal use
systems, there appear to be home systems, I think there are two
themes:

more devices on SCSI host adapter than ATA IRQ
better performance, which I think can be attributed to higher
performance mechanics of SCSI drives PLUS NCQ.
SCSI drives cost much more, up to 3X compared to 'equivalent' ATA
drives

So if SATA drives have NCQ, why not implement FORMAT UNIT and other
SCSI-only commands?

SATA NCQ isn't SCSI tagged command queuing. So adding FORMAT UNIT doesn't
produce a SCSI drive. FORMAT UNIT rewrites the sectors on the disk, even
resizing them. There is support for variable sector sizes in ATA7, but the
drives are only just getting there.

That would seem to be a way to address SCSI economics, while giving
people more choices.

Enterprise drives often have one CPU dedicated to the servo, and another one
for everything else. ATA drives usually have one CPU for everything.
That's just one example of construction differences that contribute to the
cost delta.

Sure. But that extra CPU can't account for all of the manufacturing
cost difference,assuming that it's probably an 8-bit part that costs
say $2 or so.

Quote:

Phil

great name, by the way. (You have to assume that Winey is a 'nym.)

Folkert Rienstra
Guest

Posted: Wed Sep 29, 2004 2:54 am Post subject: Re: SATA drive on SCSI bus-more questions about SAS

"Phil Barila" <PBarila@Barila.com> wrote in message news:sO-dnfY1XJB2ecXcRVn-vw@4dv.net

Quote:

"Winey" <NOSPAMME@no-one-here.com> wrote in message news:garel05qcger2ul6ii6l2u6u5r0omd3jn6@4ax.com...
On Thu, 23 Sep 2004 23:33:40 -0600, "Phil Barila" <PBarila@Barila.com> wrote:
"Winey" NOSPAMME@no-one-here.com> wrote in message news:8ls4l0p4i7b313ka1ruf3mpoiph6s4num4@4ax.com...

[snip]

A while ago I read a posting about SATA and SCSI that said, AFAIR,
that SAS would also allow attachment of SATA drives. Somehow. What
do you think?

SAS does indeed allow attachment of SATA devices. The converse isn't true,
you can't attach a SAS drive to a SATA HBA.

I'm curious about this. Is there a URL that goes through these
issues, and explains what the benefits are of doing SATA on a SAS-host
adapter? Can SATA and S-SCSI devices be mixed on the same HA?

Absolutely.

Only if that HBA has the necessary logic (expander) on board.

Quote:

It's been demonstrated several times since April.

I don't know when it's commercially available.

Right.

Quote:

Google for:
SAS SATA demo.

Here's some shameless self-promotion:

http://www.seagate.com/cda/newsinfo/newsroom/releases/article/0,1121,2074,00.html

Also, picking up on your other posting about the mechanical
differences between SCSI and ATA, if I can attach SATA to a SAS host
adapter, why not implement all the SCSI commands, with an autosense
for the type of host adapter?

SATA attachment to SAS puts all the burden on the SAS HA.

Self evident.

Quote:

The SATA drive doesn't even know it's attached to a SAS HA.

Because of the expander.

Quote:

Another selfevident. But both are "tagged command queing".
Even ATA's Queued Feature set has tagged command queueing.

Quote:

So adding FORMAT UNIT doesn't produce a SCSI drive.

Whatever that has got to do with queueing.

Quote:

FORMAT UNIT rewrites the sectors on the disk, even resizing them. There is sup-
port for variable sector sizes in ATA7, but the drives are only just getting there.

Odd. Usually it is the spec that trails behind.
Are you referring to word 106 in the identify sector?
That is still based on 512 byte logical sectors (blocks).
'Large Physical Sectors' are drive internal to increase Format efficiency.

Quote:

That would seem to be a way to address SCSI economics, while giving
people more choices.

Enterprise drives often have one CPU dedicated to the servo, and another one
for everything else. ATA drives usually have one CPU for everything.
That's just one example of construction differences that contribute to the
cost delta.

Phil

Phil Barila
Guest

Posted: Mon Oct 04, 2004 6:05 am Post subject: Re: SATA drive on SCSI bus-more questions about SAS

"Winey" <NOSPAMME@no-one-here.com> wrote in message
news:bb1il0pr6iqdj5l4iighupi0ldg6od0q9n@4ax.com...

Quote:

On Mon, 27 Sep 2004 22:13:23 -0600, "Phil Barila" <PBarila@Barila.com
wrote:

SATA attachment to SAS puts all the burden on the SAS HA. The SATA drive
doesn't even know it's attached to a SAS HA.

Oh. Does the SAS HA assign a SCSI ID to the SATA drive? LUN even?

Google is your friend. There is a plethora of info out there.

As far as the OS is concerned, it's got whatever the HBA says it does. If
the HBA says it's got 5 LUNs, the OS doesn't know any different.

Quote:

Enterprise drives often have one CPU dedicated to the servo, and another
one
for everything else. ATA drives usually have one CPU for everything.
That's just one example of construction differences that contribute to
the
cost delta.

Sure. But that extra CPU can't account for all of the manufacturing
cost difference,assuming that it's probably an 8-bit part that costs
say $2 or so.

I didn't say it's the only difference. And fi it is present, it isn't an
8-bit part.

Phil
--
Philip D. Barila Windows DDK MVP
Seagate Technology LLC
(720) 684-1842
As if I need to say it: Not speaking for Seagate.
E-mail address is pointed at a domain squatter. Use reply-to instead.

Phil Barila
Guest

Posted: Mon Oct 04, 2004 6:11 am Post subject: Re: SATA drive on SCSI bus-more questions about SAS

"Ron Reaugh" <rondashreaugh@att.net> wrote in message
news:h566d.641100$Gx4.490645@bgtnsc04-news.ops.worldnet.att.net...

Quote:

"Phil Barila" <PBarila@Barila.com> wrote in message news:sO-

Enterprise drives often have one CPU dedicated to the servo,

Which current Seagate models have two CPUs onboard?

I'm not going to answer that. Draw what conclusions you'd like, but I
assure you that 50% of you will be wrong. :-)

Phil
--
Philip D. Barila Windows DDK MVP
Seagate Technology LLC
(720) 684-1842
As if I need to say it: Not speaking for Seagate.
E-mail address is pointed at a domain squatter. Use reply-to instead.

Phil Barila
Guest

Posted: Mon Oct 04, 2004 6:15 am Post subject: Re: SATA drive on SCSI bus-more questions about SAS

"Folkert Rienstra" <see_reply-to@myweb.nl> wrote in message
news:2ru9rhF1eujjmU3@uni-berlin.de...

Quote:

"Phil Barila" <PBarila@Barila.com> wrote in message
news:sO-dnfY1XJB2ecXcRVn-vw@4dv.net
"Winey" <NOSPAMME@no-one-here.com> wrote in message
news:garel05qcger2ul6ii6l2u6u5r0omd3jn6@4ax.com...
On Thu, 23 Sep 2004 23:33:40 -0600, "Phil Barila" <PBarila@Barila.com
wrote:
"Winey" NOSPAMME@no-one-here.com> wrote in message
news:8ls4l0p4i7b313ka1ruf3mpoiph6s4num4@4ax.com...

[snip]

A while ago I read a posting about SATA and SCSI that said, AFAIR,
that SAS would also allow attachment of SATA drives. Somehow.
What
do you think?

SAS does indeed allow attachment of SATA devices. The converse
isn't true,
you can't attach a SAS drive to a SATA HBA.

I'm curious about this. Is there a URL that goes through these
issues, and explains what the benefits are of doing SATA on a SAS-host
adapter? Can SATA and S-SCSI devices be mixed on the same HA?

Absolutely.

Only if that HBA has the necessary logic (expander) on board.

Are you aware of any SAS HBAs that don't have an embedded SATA bridge? I'm
not being a smart-aleck, I'm not aware of any, but I haven't seen any up
close and personal, so my exposure is severely limited. If you know, and
can tell, please educate us all.

Phil
--
Philip D. Barila Windows DDK MVP
Seagate Technology LLC
(720) 684-1842
As if I need to say it: Not speaking for Seagate.
E-mail address is pointed at a domain squatter. Use reply-to instead.

Tim Keating
Guest

Posted: Mon Nov 22, 2004 10:45 am Post subject: Re: SATA drive on SCSI bus

On Thu, 23 Sep 2004 00:43:48 +0200, "Folkert Rienstra"
<see_reply-to@myweb.nl> wrote:

Quote:

"Keve Nagy" <no_spam@poliod.hu> wrote in message news:2rctbdF18cgqvU1@uni-berlin.de
Good Morning Newsgroup,

I was wondering, what makes SCSI better than IDE/ATA/ATAPI/SATA.

The application.

More precisely what I really wonder about is if the SCSI controller and
the bus alone (not including the devices) can make the difference in
range of raliability, speed and services.

Nothing much.
SCSI has a parity signal as an extra precaution against all data corruption
for which ATA has CRC but only for UDMA and only userdata transfers.

CRC was added on top of parity check for devices running data phases
at Ultra 3/160 speeds.

Additional data integrity items were added to Ultra 320, like domain
validation (actually calibrates cable config between host adapter and
each device.).

Quote:

SATA is the ATA equivalent of Serial Attached Scsi (SAS).
SATA maybe just ATA with a serial convertor.
ATAPI is basically SCSI over ATA.

No.. The original ATAPI is the old IBM AT ST506/Western digital
controller interface connected via an extension of a IBM AT 16 bit
parallel IO bus. Not even close to the SCSI standards..

Note: Subsequent versions improved upon this design by adding DMA,
faster data rates, CRC and extra controlled register definitions..

Folkert Rienstra
Guest

Posted: Tue Nov 23, 2004 5:03 am Post subject: Re: SATA drive on SCSI bus

"Tim Keating" <NotForJunkEmail@directinternet11.com1> wrote in message news:p8u2q055csbave4bc0kri5sk5d3gtpic2k@4ax.com

Quote:

On Thu, 23 Sep 2004 00:43:48 +0200, "Folkert Rienstra" <see_reply-to@myweb.nl> wrote:
"Keve Nagy" <no_spam@poliod.hu> wrote in message news:2rctbdF18cgqvU1@uni-berlin.de
Good Morning Newsgroup,

I was wondering, what makes SCSI better than IDE/ATA/ATAPI/SATA.

The application.

More precisely what I really wonder about is if the SCSI controller and
the bus alone (not including the devices) can make the difference in range
of raliability, speed and services.

Nothing much.
SCSI has a parity signal as an extra precaution against all data corruption
for which ATA has CRC but only for UDMA and only userdata transfers.

CRC was added on top of parity check for devices running data phases
at Ultra 3/160 speeds.

Additional data integrity items were added to Ultra 320,

like domain validation

That was already in Ultra160 too.

Quote:

(actually calibrates cable config between host adapter and each device.)

Actually avoids negotiating non-viable speeds by doing test transfers.

Quote:

SATA is the ATA equivalent of Serial Attached Scsi (SAS).
SATA may be just ATA with a serial convertor.
ATAPI is basically SCSI over ATA.

No..

Yup.

Quote:

The original ATAPI is the old IBM AT ST506/Western digital
controller interface connected via an extension of a IBM AT 16 bit
parallel IO bus.

Nonsense, ATAPI (AT Attachment with Packet Interface) didn't exist
until the 4th revision of the ATA standard.

Quote:

Not even close to the SCSI standards..

Actually, ATAPI is covered under the SCSI MMC standard.
And who said anything about parallel SCSI? I said "SCSI over ATA".

Quote:

Note: Subsequent versions improved upon this design by adding DMA,
faster data rates, CRC and extra controlled register definitions..

You must be a real SCSI zealot to know such details as Double Transi-
tion Data phases and the covering of CRC but not have a single clue on
ATA(PacketInterface).

Tim Keating
Guest

Posted: Thu Nov 25, 2004 5:41 pm Post subject: Re: SATA drive on SCSI bus

On Tue, 23 Nov 2004 01:03:10 +0100, "Folkert Rienstra"
<see_reply-to@myweb.nl> wrote:

Quote:

"Tim Keating" <NotForJunkEmail@directinternet11.com1> wrote in message news:p8u2q055csbave4bc0kri5sk5d3gtpic2k@4ax.com
On Thu, 23 Sep 2004 00:43:48 +0200, "Folkert Rienstra" <see_reply-to@myweb.nl> wrote:
"Keve Nagy" <no_spam@poliod.hu> wrote in message news:2rctbdF18cgqvU1@uni-berlin.de
Good Morning Newsgroup,

I was wondering, what makes SCSI better than IDE/ATA/ATAPI/SATA.

The application.

More precisely what I really wonder about is if the SCSI controller and
the bus alone (not including the devices) can make the difference in range
of raliability, speed and services.

Nothing much.
SCSI has a parity signal as an extra precaution against all data corruption
for which ATA has CRC but only for UDMA and only userdata transfers.

CRC was added on top of parity check for devices running data phases
at Ultra 3/160 speeds.

Additional data integrity items were added to Ultra 320,

like domain validation

That was already in Ultra160 too.

(actually calibrates cable config between host adapter and each device.)

Actually avoids negotiating non-viable speeds by doing test transfers.

Ultra -320 performs a cable (calibration) deskewing function used
for Paced data (fast-160 x16bit) transfers.

A quick summary can be found at..
http://www.t10.org/ftp/t10/document.00/00-133r2.pdf

The calibration values are retained in each device for each I_T nexus
and is used to for all subsequent Paced data transfers.

Quote:

SATA is the ATA equivalent of Serial Attached Scsi (SAS).
SATA may be just ATA with a serial convertor.
ATAPI is basically SCSI over ATA.

No..

Yup.

The original ATAPI is the old IBM AT ST506/Western digital
controller interface connected via an extension of a IBM AT 16 bit
parallel IO bus.

Nonsense, ATAPI (AT Attachment with Packet Interface) didn't exist
until the 4th revision of the ATA standard.

ATAPI is still just an extension of the IBM AT's old 16 bit IO bus.
Not even close to any the parallel SCSI standards..

Quote:

Not even close to the SCSI standards..

Actually, ATAPI is covered under the SCSI MMC standard.
And who said anything about parallel SCSI? I said "SCSI over ATA".

Huh? Just what does the SCSI Multi-Media Command definitions
(CD-ROM, etc) have to do with ATA???

You've got your ANSI committees (T10, T13) mixed up..

Quote:

Note: Subsequent versions improved upon this design by adding DMA,
faster data rates, CRC and extra controlled register definitions..

You must be a real SCSI zealot to know such details as Double Transi-
tion Data phases and the covering of CRC but not have a single clue on
ATA(PacketInterface).

You are one very confused puppy..

Eric Gisin
Guest

Posted: Thu Nov 25, 2004 8:34 pm Post subject: Re: SATA drive on SCSI bus

"Tim Keating" <NotForJunkEmail@directinternet11.com1> wrote in message
news:ucfbq0p2om7kiv34i071ek79up7s64bhp7@4ax.com...

Quote:

On Tue, 23 Nov 2004 01:03:10 +0100, "Folkert Rienstra"
see_reply-to@myweb.nl> wrote:

Actually, ATAPI is covered under the SCSI MMC standard.
And who said anything about parallel SCSI? I said "SCSI over ATA".

Huh? Just what does the SCSI Multi-Media Command definitions
(CD-ROM, etc) have to do with ATA???

You've got your ANSI committees (T10, T13) mixed up..

Clueless. Have you even have read MMC?

Folkert Rienstra
Guest

Posted: Fri Nov 26, 2004 4:23 am Post subject: Re: SATA drive on SCSI bus

"Tim Keating" <NotForJunkEmail@directinternet11.com1> wrote in message news:ucfbq0p2om7kiv34i071ek79up7s64bhp7@4ax.com

Quote:

On Tue, 23 Nov 2004 01:03:10 +0100, "Folkert Rienstra" <see_reply-to@myweb.nl> wrote:
"Tim Keating" <NotForJunkEmail@directinternet11.com1> wrote in message news:p8u2q055csbave4bc0kri5sk5d3gtpic2k@4ax.com
On Thu, 23 Sep 2004 00:43:48 +0200, "Folkert Rienstra" <see_reply-to@myweb.nl> wrote:
"Keve Nagy" <no_spam@poliod.hu> wrote in message news:2rctbdF18cgqvU1@uni-berlin.de
Good Morning Newsgroup,

I was wondering, what makes SCSI better than IDE/ATA/ATAPI/SATA.

The application.

More precisely what I really wonder about is if the SCSI controller and
the bus alone (not including the devices) can make the difference in range
of raliability, speed and services.

Nothing much.
SCSI has a parity signal as an extra precaution against all data corruption
for which ATA has CRC but only for UDMA and only userdata transfers.

CRC was added on top of parity check for devices running data phases
at Ultra 3/160 speeds.

Additional data integrity items were added to Ultra 320,

like domain validation

That was already in Ultra160 too.

(actually calibrates cable config between host adapter and each device.)

Actually avoids negotiating non-viable speeds by doing test transfers.

Ultra -320 performs a cable (calibration) deskewing function used
for Paced data (fast-160 x16bit) transfers.

But that has nothing got to do with Domain Validation other than that they
both occur at negotiation time. The term Domain Validation doesn't even
turn up in that document referred to below ....

Quote:

A quick summary can be found at..
http://www.t10.org/ftp/t10/document.00/00-133r2.pdf

... nor does Paced data transfers.

Quote:

The calibration values are retained in each device for each I_T nexus
and is used to for all subsequent Paced data transfers.

SATA is the ATA equivalent of Serial Attached Scsi (SAS).
SATA may be just ATA with a serial convertor.
ATAPI is basically SCSI over ATA.

No..

Yup.

The original ATAPI is the old IBM AT ST506/Western digital
controller interface connected via an extension of a IBM AT 16 bit
parallel IO bus.

Nonsense, ATAPI (AT Attachment with Packet Interface) didn't exist
until the 4th revision of the ATA standard.

ATAPI is still just an extension of the IBM AT's old 16 bit IO bus.
Not even close to any the parallel SCSI standards..

Right, which isn't remotely close to what you said before.
Nor did I imply that is was IDE converted to parallel SCSI.
Of all the SCSI standards it is actually the closest to the Parallel
SCSI standard.

And here is what the Scope of SPI-4 says:

Figure 1 - General Structure of SCSI

At the time this standard was generated examples of the SCSI general structure included:
Physical Interconnects:

Quote:

AT Attachment with Packet Interface Extension [NCITS.317-1998]
Fibre Channel Arbitrated Loop-2 [T11/1133D]

Fibre Channel - Physical and Signaling Interface [X3.230-1994]
High Performance Serial Bus [IEEE 1394-1995]
SCSI Parallel Interface - 2 [X3.302-1998]
SCSI Parallel Interface - 3 [NCITS.336-200x]
SCSI Parallel Interface - 4 [This standard]
Serial Storage Architecture Physical Layer 1 [X3.293-1996]
Serial Storage Architecture Physical Layer 2 [NCITS.307-1998]
.....

and

"The term SCSI is used wherever it is not necessary to distinguish between
the versions of SCSI".

Quote:

Not even close to the SCSI standards..

Actually, ATAPI is covered under the SCSI MMC standard.
And who said anything about parallel SCSI? I said "SCSI over ATA".

Huh? Just what does the SCSI Multi-Media Command definitions
(CD-ROM, etc) have to do with ATA???

Everything with ATAPI, oh mighty zealot one.

Quote:

You've got your ANSI committees (T10, T13) mixed up..

LOL, nice try. Try again.

Quote:

And you are an obvious TROLL. A wannabe zealot TROLL at that.

Peter H.
Guest

Posted: Sun Nov 28, 2004 11:42 am Post subject: Re: SATA drive on SCSI bus

Quote:

The original ATAPI is the old IBM AT ST506/Western digital
controller interface connected via an extension of a IBM AT 16 bit
parallel IO bus.

Nonsense, ATAPI (AT Attachment with Packet Interface) didn't exist
until the 4th revision of the ATA standard.

ATAPI is still just an extension of the IBM AT's old 16 bit IO bus.
Not even close to any the parallel SCSI standards..

Quote:

ATAPI is simply putting all the functionality of a WD-1003 controller on the
drive.

The controller in the AT was a WD-1003.

(The controller in the XT was an IDE bus variation of Xebec's SASI controller,
with the function of the "host adapter" replacing the function of the SASI bus
responder ... otherwise, the XT controller used almost all of the same parts,
and much of the same firmware, as the Xebex 1410 controller).

	PC Hardware Forum Index -> SCSI	All times are GMT Goto page Previous 1, 2, 3 Next
Page 2 of 3

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum

Electronics VoIP DSP
New Topics php BB