[cups.general] IPP network overhead per-job

[Michael R Sweet]

Sebastian James wrote:
> ...
 > Would anyone know if the "add multiple files to a cups job" method
> would be expected to reduce network overhead per-job? Or would it be
 > the same as sending each file in a single cups job?

For CUPS 1.3 and higher using raw queues, doing a single job with
multiple files will result in fewer round-trip requests - a minimum
of 3 + N instead of 3N, where N is the number of files.  The total
number of requests depends on how long it takes to print a job (the
backend will poll the remote server periodically to see when the job
completes).

For a non-raw queue there is no difference - the backend reads all of
the print data for the job and submits a single large file with all
of the data.

You can also (in CUPS 1.3 and higher) tell CUPS to gzip each of the
print files using the compression option in the URI, e.g.:

     /usr/sbin/lpadmin -p rwssl4 -E -v 
"ipp://209.119.222.222:631/?encryption=always&compression=gzip" -m raw

The receiving system/printer needs to support gzip compression, of
course (CUPS 1.2 and higher does, as do some network printers).

Finally, you can use the waitjob and waitprinter options to eliminate
the post-submission checks so that you start sending the next job
immediately:

     /usr/sbin/lpadmin -p rwssl4 -E -v 
"ipp://209.119.222.222:631/?encryption=always&compression=gzip&waitjob=false&waitprinter=false" 
-m raw

Note that the waitjob option *can* result in lost print data since we
don't retry the job if it fails to print...

-- 
______________________________________________________________________
Michael R Sweet                        Senior Printing System Engineer