AIT Drives Under USB?

[Will]

Has anyone been successful in using any flavor of AIT tape drive under USB,
mapped through IDE? I know Sony makes an IDE version of the drive, and it's
common these days to find cases that will map IDE/EIDE hard drives into USB
2.0. I assume those same casese could map an AIT IDE drive to the USB 2.0
bus, but I have no idea if the Sony tape drivers would work well under
Windows 2000/XP when using the tape drive through the USB 2.0 interface.

 

[Maxim S. Shatskih]

Do you mean the USB-to-IDE adapter boxes? If yes - they usually support
only a limited subset of SCSI, so I have doubts that the tape will work. You
can try though.

 

[Will]

What would SCSI have to do with this if the tape device has a native IDE
interface?

 

[Maxim S. Shatskih]

Any IDE tape drive is not IDE, it is ATAPI, which means - SCSI commands
over the IDE cable.

Same is true for CD/DVD drives.

Same is true for USB storage, with the additional limitation of the USB
storage spec supporting a limited set of SCSI commands.

 

[Rod Speed]

Any IDE tape drive is not IDE, it is ATAPI, which
means - SCSI commands over the IDE cable.

Wrong, as always.

Same is true for CD/DVD drives.

Wrong, as always.

Same is true for USB storage, with the additional limitation of
the USB storage spec supporting a limited set of SCSI commands.

Wrong, as always.

 

[Maxim S. Shatskih]

Any IDE tape drive is not IDE, it is ATAPI, which

Wrong, as always.

Trolling? Go read the specs on www.t13.org

 

[Rod Speed]

Trolling?

Pig ignorant wanking in your case.

Go read the specs on www.t13.org

Been there, done that, long ago now.

Doesnt say that.

 

[Folkert Rienstra]

Utterly clueless, as always.

Wrong, as always.


Wrong, as always.


Wrong, as always.

 

[Folkert Rienstra]

Trolling?

Corse it is. What did you do to him to earn it's attention.

Go read the specs on www.t13.org

Pointless. Never showed any signs that it can read specs.

 

[Folkert Rienstra]

Has anyone been successful in using any flavor of AIT tape drive under USB,
mapped through IDE? I know Sony makes an IDE version of the drive, and it's
common these days to find cases that will map IDE/EIDE hard drives into USB 2.0.

1394 to AT Attachment - Tailgate,
specifies the protocol for passing ATA
and ATAPI commands over the 1394 bus.

I would expect the USB to IDE protocol to have a similar extension.
Unfortunately T13 does't seem to have been involved with that too.

USB + tailgate get's this in Google though:

http://www.cooldrives.com/dubayusb205o.html
"USB-MS-2BW is ideally suited for tailgate interface -applications for
removable-media drives CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM,
tape drive, and hard disk drive.
It allows IDE drives (Master and Slave) to be connected to USB 2.0 serial
bus in a plug-and-play fashion. Both ATA and ATAPI devices are supported
using the same firmware.

I assume those same cases could map an AIT IDE drive to the USB 2.0
bus, but I have no idea if the Sony tape drivers would work well under
Windows 2000/XP when using the tape drive through the USB 2.0 interface.

You would have to ask Sony whether the extension driver has an API interface to USB too.

 

[Maxim S. Shatskih]

I would expect the USB to IDE protocol to have a similar extension.

USBSTOR protocol supports only some subset of SCSI commands, not all of them.
The details are in the spec on www.usb.org

 

[Folkert Rienstra]

USBSTOR protocol supports only some subset of SCSI commands,

Which limits it to the use of what? Which SCSI classes are excluded?

not all of them.

And even if it did, what would that mean for ATA/ATAPI devices.
And even if it did, the bridge doesn't necessarily translate all of them
into ATA/ATAPI commands.

The details are in the spec on www.usb.org

But nothing I could find about ATA/ATAPI, at least not without reading it
word for word.

 

[Maxim S. Shatskih]

USBSTOR protocol supports only some subset of SCSI commands,

Which limits it to the use of what? Which SCSI classes are excluded?

Don't remember. I only remember that the subset has some name, and is defined
in main SCSI specs. USBSTOR spec claims to only implement this subset.

 

[Eric Gisin]

Don't remember. I only remember that the subset has some name, and is defined
in main SCSI specs. USBSTOR spec claims to only implement this subset.

There are six subclasses of the USB mass storage class.
There is one that uses RBC instead of SBC, which Microsoft does not support.
Otherwise, SPC, SBC and MMC seem to be preferred.

Here as a good USB developer page: http://www.lvr.com/usb.htm .

=== USB Storage Classes ===

SubClass Code Command Block Specification Comment

01h Reduced Block Commands (RBC) T10 Project 1240-D
Typically, a Flash device uses RBC command blocks.
However, any Mass Storage device can use RBC command blocks.

02h SFF-8020i, MMC-2 (ATAPI)
Typically, a C/DVD device uses SFF-8020i or MMC-2 command blocks for its Mass Storage interface.

03h QIC-157 Typically, a tape device uses QIC-157 command blocks.

04h UFI Typically a floppy disk drive (FDD) device.

05h SFF-8070i Typically, a floppy disk drive (FDD) device uses
SFF-8070i command blocks. However, an FDD device can be in another subclass (for example, RBC)
and other types of storage devices can belong to the SFF-8070i subclass.

06h SCSI transparent command set.

07h - FFh Reserved for future use.

 

[Folkert Rienstra]

Which presumably is the Host Program Interface, not necessarily what
travels over the USB cable.
Apparently USB leaves it to the OS on how to implement this interface.

Don't remember.

I only remember that the subset has some name,
Obviously.

and is defined in main SCSI specs.

Can't be. SCSI specs don't mention USB at all.
(Neither does USB spec mention SCSI at all for that matter).

USBSTOR spec claims to only implement this subset.

A subset of SCSI Primary Commands (SPC) presumably.

I don't think that that is necessarily a problem.
IINM, you need extension drivers for SCSI as well if
you want to use other than disk drives.

 

[Maxim S. Shatskih]

I only remember that the subset has some name,

Obviously.

The name is Reduced Block Commands (RBC) (thanks to Eric).

Can't be. SCSI specs don't mention USB at all.

They mention Reduced Block Commands.

(Neither does USB spec mention SCSI at all for that matter).

The USBSTOR spec references the T10 specs (SCSI).

IINM, you need extension drivers for SCSI as well if
you want to use other than disk drives.

In Windows starting from XP, you can use CD/DVD burners off USB (and also off
1394).
In w2k you can only use the disk storage off USB, not the burner.

 

[Folkert Rienstra]

The name is Reduced Block Commands (RBC) (thanks to Eric).


They mention Reduced Block Commands.

Indeed they do.
I expected the other than SPC specs to be supplemental to SPC
and since you mentioned a (=one) spec I assumed you meant SPC.
As it turns out RBC also includes the necessary SPC commands
for RBC device implementation.

The USBSTOR spec

Is there a home for that spec?

references the T10 specs (SCSI).

Right, as I said, the USB spec leaves that to the OS implementation.

Any ideas about the tail end?

In Windows starting from XP, you can use CD/DVD burners off USB (and also off
1394).

In w2k you can only use the disk storage off USB, not the burner.

No backwards compatibility with using XP drivers?

 

[Maxim S. Shatskih]

For the spec, go to

http://www.usb.org/developers/devclass_docs/

and look at "Mass Storage". The spec itself is a PDF, and I do not want to
give the URL leading to a PDF directly.

Right, as I said, the USB spec leaves that to the OS implementation.

Any ideas about the tail end?

Sorry, the SCSI/RBC commands are going across the USB wire. What "OS
implementation" you're speaking about, considering that the spec also defines
the _device side_, and not only the host side?

The main purpose of the spec is to define the on-the-wire traffic (so that the
driver writers and hardware developers will have some common guideline), all of
the rest is really OS-dependent.

 

[Folkert Rienstra]

For the spec, go to

http://www.usb.org/developers/devclass_docs/

Ah. That one was well hidden in the developers area.

and look at "Mass Storage". The spec itself is a PDF, and I do not want to
give the URL leading to a PDF directly.


Sorry, the SCSI/RBC commands are going across the USB wire.

For the USB mass storage driver, presumably fixed disk.
Not necessarily for other drive(r)s, e.g. for Floppy disk it
is a flavor of ATAPI, for tape it is (or can be) QIC.

That's an extra layer of protocol that sits on top of the USB driver.
In theory the same extra layer sits on top of the device-side USB interface.
But with USB storage bridges the HW side is IDE/ATA(PI), not SCSI.
So there should be some description for the (USB)SCSI to IDE translation,
an USB tailgate.

What "OS implementation" you're speaking about, considering that the
spec also defines the _device side_,

That's the USB wire side, not the host or device HW interface side.

and not only the host side?

The Host side protocol is in an extra software layer that sits on top of
the USB driver.
The specs for that are defined by the USB Device Working Group.

The main purpose of the spec is to define the on-the-wire traffic

That's USB.

(so that the
driver writers and hardware developers will have some common guideline),

That's not USB. That's the device class driver.