SCSI FAQ 41-44

COMP.PERIPHS.SCSI FAQ


Contents:



41) What is the difference between SCSI-2 and SCSI-3?

ANSWER From: excerpts of postings by Jeff Stai and others: (Mohit K Goyall - goyall@utdallas.edu), (Andrew E. Lowman - lowman@arizona.edu)

Are SCSI-3 hard drives and/or controllers available yet?

Allegedly. Previous postings have said "I heard that SCSI-3 has been standardized," but I haven't seen anything firm about it. I've seen controllers advertised by JDR Microdevices and some cheap clones; the Quantum "Empire" drives are also advertised as SCSI-3 by some mail order vendors. Seagate and IBM call their fastest drives (probably comparable in speed to the Quantums, if not faster) "Wide SCSI-2."

That's a misnomer. See below.

What is the difference between SCSI-3 and Fast & Wide SCSI-2?

Wide SCSI-2 required two cables to do 16 bit wide transfers. SCSI-3 defined a single cable, single REQ/ACK 16 bit wide transfer. The reason you are hearing 16-bit single cable being called SCSI-3 is that they CAN. The fact that single cable 16-bit has been around for a while just shows you how much the standardization process lags behind the real world.

SCSI-3 is really a family of standards. SCSI was broken up from a single document into different layers and command sets. This was done to allow for different physical transport layers (like fibre channel and SSA) to be defined, and to allow for smaller "bite-sized" projects that maybe get done a little faster ;-)

The family includes the following members with TLAs:

SCSI-3 Parallel Interface (SPI)
Defines the mechanical, timing, phases, and electrical parameters of the parallel cable we all know and love. Some of the electrical and cable parameters are tightened/improved over SCSI-2.

SCSI-3 Interlock Protocol (SIP)
Defines the messages and how the phases are invoked. No real change from SCSI-2, except for some new messages.

SCSI-3 Architectural Model (SAM)
In a nutshell, defines a common set of functions and services and definitions for how a physical transport properly gets commands, data, and status exchanged between two devices, complete with error handling and queueing.

SCSI-3 Primary Commands (SPC)
All of the commands executed by any and all SCSI devices, like REQUEST SENSE and INQUIRY, etc.

SCSI-3 Block Commands (SBC)
Disk commands.

SCSI-3 Stream Commands (SBC)
Tape commands.

SCSI-3 Controller Commands (SCC)
RAID box commands.

SCSI-3 Multimedia Commands (MMC)
For CDROMS etc.

SCSI-3 Fibre Channel Protocol (FCP)
SCSI commands over gigabit Fibre Channel.

SCSI-3 Serial Bus Protocol (SBP)
SCSI commands over IEEE 1394 High Speed Serial Bus (Apple's "Firewire").

SCSI-3 Serial Storage Protocol (SSP)
SCSI commands over SSA.

[whew.]

After perusing the latest issue of Computer Shopper, I came away with the impression that companies are calling F&W scsi-2 hd's SCSI-3. Is this an incorrect assumption, or is F&W SCSI-2 known as SCSI-3? Is this really mostly marketing hype?

Actually, there is something to that. TECHNICALLY, what is out there is often a hybrid: SCSI-3 "SPI" silicon with some other hodgepodge of SCSI-3 proposals, all mixed in with SCSI-2 stuff.

An earlier posting said that the Quantum Empire ("SCSI-3") drives contain some commands from the SCSI-3 command set, and Adaptec suggested a specific setting on its 2940W controller to work properly with the drive.

I understand there are some drives with proposed SCSI-3 command features. These are mostly in the MODE SELECT and in error codes, as I recall. Perhaps someone who knows more about this could elaborate?

Note also that the major players (like DC Drives) don't have any "SCSI-3" stuff advertised; only JDR and some cheap clones are promoting it. Besides, Wide SCSI-2 has yet to really catch on (mostly because only a few drives are fast enough to take advantage of it).

There is no "wide SCSI-2" because that would mean two cables. Single cable wide SCSI has always been SCSI-3, it just took too d*** long to get into a standard!-)


42) Is SYNCHRONOUS faster than ASYNCHRONOUS?

(see next entry)


43) Is the 53C90 Faster than spec?

ANSWER From: kstewart@ncr-mpd.FtCollins.NCR.COM (Ken Stewart)

I've seen a few comments about our 54C90 being faster than spec. While I doubt the author was really complaining (I got twice as much as I paid for--sure makes me mad ;-) I'd like to explain the situation.

Along the way, I'll also show that asynchronous is faster on short cables, while synchronous is faster on long cables. The cross-over point occurs somewhere around six feet--assuming that you have our 53C90 family devices at both ends of the cable. The reason has to do with the propagation delay of the cable; the turn around time of the silicon; and the interlocked nature of the asynchronous handshake.

  1. We have measured propagation delays from various cables and found an average of 1.7 nanoseconds per foot, which is roughly 5.25 ns per meter.

  2. The turn-around time is the amount of time the SCSI chip takes to change an output in response to an input. If REQ is an input then ACK is an output. Or if ACK is an input then REQ is an output. Typical turn-around time for the 53C90 is 40 nanoseconds.

  3. The asynchronous transfer uses an interlocked handshake where a device cannot do the next thing until it receives positive acknowledgment that the other device received the last thing.

        First REQ goes true                   /* driven by Target */
        then ACK is permitted to go true      /* driven by Initiator */
        then REQ is permitted to go false
        then ACK is permitted to go false
    

Thus we have four "edges" propagating down the cable plus 4 turn-around delays. Asynchronous transfer requires 55 ns setup and no hold time (paragraph in 5.1.5.1 in SCSI-1 or SCSI-2) which gives an upper speed limit around 18 MB/s. A detailed analysis (assuming 53C90 family) shows that the setup time subtracts out. This is mostly because we are running at one-third the max rate, but also because setup for the next byte can begin anytime after ACK is received true or REQ is received false, depending on who is receiving. You can either take my word for it or draw the waveforms yourself. Thus, the asynchronous transfer reduces to:

(4 * 1.7 * 1) + (4 * 40ns) = 167 ns           /* 1 foot cable */
                           = 6 MB/s
 
(4 * 5.25 * 6) + (4 * 40ns) = 286 ns          /* 6 meter cable */
                            = 3.5 MB/s
 
(4 * 5.25 * 25) + (4 * 40ns) = 685 ns         /* 25 meter cable */
                             = 1.5 MB/s

Note: cables longer than 6 meters require external differential transceivers which add delay and degrade the performance even more than indicated here.

Our simulations say that under very best conditions (fast silicon, low temperature, high voltage, zero length cable) we can expect more than 8 MB/s asynchronously. In the lab, I routinely measure 5 MB/s on 8 foot cables. So, if you were writing the data manual for this, how would YOU spec it?

The framers of the SCSI spec threw in synchronous mode to boost the performance on long cables. In synchronous mode, the sending device is permitted to send the next byte without receiving acknowledgment that the receiver actually received the last byte. Kind of a ship and pray method. The acknowledgment is required to come back sometime, but we just don't have to wait for it (handwave the offset stuff and the ending boundary conditions). In this mode any external transceivers add a time shift, but not a delay. So if you negotiate for 5 MB/s, you get 5MB/s regardless how long the cable is and regardless whether you are single-ended or differential. But you can't go faster than 5.5 MB/s, except in SCSI-2. Synchronous mode does have a hold time (unlike asynch) but again, setup and hold times subtract out. In SCSI-1 synchronous mode, the speed limit comes from the combined ASSERTION PERIOD + NEGATION PERIOD which is 90ns + 90ns = 180ns = 5.5 MB/s. Our 53C90 family doesn't quite hit the max, but we do guarentee 5.0 MB/s. In SCSI-2, anything above 5.0 MB/s is considered to be FAST. Here the maximum transfer rate is explicitly limited to 100 ns or 10MB/s; you don't have to read between the lines to deduce it.

Interesting tid-bit: given a SCSI-2 FAST period of 100 ns and a cable delay of 131 ns on a 25 meter cable, you can actually stack 1.31 bytes in the 8-bit cable. In FAST and WIDE SCSI you can stack 5.24 bytes in this copper FIFO. Hummm...


44) What are the jumpers on my Conner drive?

ANSWER From: ekrieger@quasar.hacktic.nl (Eric Krieger) and Embellishment from: Henrik Stahl (f92-hst@nada.kth.se)

QUICK INSTALLATION GUIDE

SCSI

Most SCSI host adapters are compatible with Conner drives. Software drivers and installation instructions are provided with the host adapter.

The drives are shipped with SCSI ID set to 7. To select a different ID refer to the following:

Table A                  Table B

ID   E-1  E-2  E-3       ID   E2   E3   E4
==================       ==================
0    out  out  out       0    out  out  out
1    in   out  out       1    in   out  out
2    out  in   out       2    out  in   out
3    in   in   out       3    in   in   out
4    out  out  in        4    out  out  in
5    in   out  in        5    in   out  in
6    out  in   in        6    out  in   in
7    in   in   in        7    in   in   in

Parity is always ENABLED on the CP3200,CP30060,CP30080,CP30100, CP 30200, CP 3500, CP 3360, CP 30540 and CP 31370.

For the CP 340, jumper E-1 to disable parity.

All other models, jumper E-4 to disable parity.

SCSI drive parameters:

  Model       Hds    Cyl    Sec   Table   LED
  ===========================================
  CP2020      2      642    32     A      n/a
  CP340       4      788    26     B      1
  CP3020      2      622    33     A      1
  CP3040      2     1026    40     A      1
  CP3180      6      832    33     A      1
  CP3100      8      776    33     A      1
  CP30060     2     1524    39     A      2
  CP30080     4     1053    39     A      2
  CP30100     4     1522    39     A      2
  CP30200     4     2119    49     A      2
  CP3200      8     1366    38     A      2
  CP3360      8     1806    49     A      2
  CP3540      12    1806    49     A      2
  CP 30080E   2     1806   46      AA    C/E
  CP 30170E   4     1806   46      AA    C/E
  CP 30540    6     2249   59-89   AA     B
  CP 31370    14    2094   59-95   AA     B
 
LED 1               LED 2
J-4  Pin 1 = +      J-1 Pin 3 = +
     Pin 2 = -          Pin 4 = -

On the CP 31370, jumper E5 enables termination. Default is termination on. It may be the same jumper for other models.


Please feel free to mail us with suggestions, comments, and help. Give us a few days to respond,this account is not constantly active. Thanx.