SDP Mail Archive: Re: SDP> Suggestions for discussion at SDP session next week

Re: SDP> Suggestions for discussion at SDP session next week

Robert Herriot (robert.herriot@Eng.Sun.COM)
Thu, 14 May 1998 15:04:45 -0700

--=====================_10301012==_.ALT
Content-Type: text/plain; charset="us-ascii"

I just talked with Paul about the requirement for a separate data channel.

I think that Paul and I are in agreement that "separate data channel" does
not necessarily mean that an operation gets a separate channel for data in
the midst of an IPP/SDP operation. It has at least some elements of the
following.:

a) each operation opens a new socket and gets its own channel, or uses
an existing socket after the preceding operation has completed.
b) a SendDocument gets a different channel from a CreateJob because
CreateJob returns a URI which specifies a different host or port.
c) data intensive operations, such as PrintJob and Send Document use a
different port from other operations.
d) the same port is used for all operations, but the printer is able
handle connections fast enough that there will never be time-out
problems.

The important concept is that the mechanism for separate channels is handled
by a layer below IPP/SDP.

At 01:27 PM 5/14/98 , Roger K Debry wrote:
>snip ...
>
>In my conversations with Paul, he had no requirement for a separate data
>channel.
><RKD> I thought this was one of Paul's requirements. Paul ???
>
>snip ...
>
>We should define such a lower layer only if the USB and 1394
>people fail to define such a layer.
><RKD> I thought non-network attached printers were what were
><RKD> largely driving this discussion, i.e. $299 parallel port
><RKD> attached printers. I think all of the debate on TCP/IP
><RKD> is interesting, but I'd really like to see the 1394 solution
><RKD> FIRST!
><RK
>

--=====================_10301012==_.ALT
Content-Type: text/html; charset="us-ascii"

I just talked with Paul about the requirement for a separate data channel.

I think that Paul and I are in agreement that "separate data channel" does
not necessarily mean that an operation gets a separate channel for data in
the midst of an IPP/SDP operation. It has at least some elements of the following.:

    a)  each operation opens a new socket and gets its own channel, or uses
        an existing socket after the preceding operation has completed.
    b) a SendDocument gets a different channel from a CreateJob because
        CreateJob returns a URI which specifies a different host or port. 
    c) data intensive operations, such as PrintJob and Send Document use a
        different port from other operations. 
    d) the same port is used for all operations, but the printer is able
        handle connections fast enough that there will never be time-out problems.

The important concept is that the mechanism for separate channels is handled
by a layer below IPP/SDP.


At 01:27 PM 5/14/98 , Roger K Debry wrote:
>snip ...
>
>In my conversations with Paul, he had no requirement for a separate data
>channel.
><RKD> I thought this was one of Paul's requirements. Paul ???
>
>snip ...
>
>We should define such a lower layer only if the USB and 1394
>people fail to define such a layer.
><RKD> I thought non-network attached printers were what were
><RKD> largely driving this discussion, i.e. $299 parallel port
><RKD> attached printers.  I think all of the debate on TCP/IP
><RKD> is interesting, but I'd really like to see the 1394 solution
><RKD> FIRST!
><RK
>

--=====================_10301012==_.ALT--