Network Load from Printing (SAP Library - Network Integration Guide (BC-NET))

Network Load from Printing
Printing from the SAP System

The printing format and the data transmission to the printer or the spool system normally occurs in the spool work process of the SAP System. The amount of data and the destination chosen by the spool work process for the print data are both factors in network load.

The amount of data to be printed depends on:

The type of printing format (printer device type for example, HP-PCL or postscript)
The protocol of the data transmission (for example, Berkeley lpd or SAPlpd)

The route over which the data is sent depends on the access method (local or remote printing).

Access Method for Local Printing

L = Unix local printing with lp/lpr

C = local printing with an NT operating system

E = different output management system

Access Method for Remote Printing

U = Printing to an LPD host using Berkeley protocol

S = Printing to SAPlpd host using SAP protocol

F = Printing on the frontend host

Network Load when Printing

How Size of Print Jobs Affects Network Load

Small print jobs of only one page result in an overhead of approximately 100% due to the print job administration in relation to the transmitted data. Print jobs of 10 pages lower the administration overhead to approximately 10%.

If possible, ensure that print jobs are as large as possible. Avoid printing only single pages.

How Access Methods Affect Network Load

The access method that uses the Berkeley protocol transmits the print data uncompressed over the network. The network load corresponds to the number of characters to be printed and the print job administration.

The SAPlpd access method differs from the Berkeley protocol in that the print data is compressed. For full body text pages, the compression rate reaches up to 3:1.

The frontend printing access method causes a greater overhead due to processing through the connection to the SAPgui frontend. Therefore, only use this method in exceptional cases.

If possible, use access method S (SAPlpd).

How Output Devices Affect Network Load

Postscript devices generate approximately 20% overhead for body text.

HP PCL devices generate approximately 10% overhead for body text.

SAPWIN devices generate approximately 8% overhead for body text.

Output devices only have a marginal effect on network load.

Determining the Customer-Specific Network Load

In LAN environments, printing usually causes no problems regarding network load. For business locations connected through a WAN, a large amount of print data can lead to problems. To best determine the network load caused by printing, you must determine the size of the print-formatted documents. You must consider the documents that are printed most often at the relevant locations.

Proceed as follows:

Call Transaction SPAD.
Switch to change mode and choose an output device.
Choose Edit ® Debugger.
Select the option Retain print file.

After printing, the print data is stored in the /usr/sap/<SID>/<instance>/data directory. You can view the size of the data in the file system.

Measurement Results

Components when Measuring

R/3 Release 4.0B

SAPlpd Version 4.08

Conditions for Measurement Results

Only body text pages without formatting were used for the measurements. Three measurements were made for each configuration.

Measurement: 1 page (65x80) = 5200 characters
Measurement: 10 pages (65x80) = 52000 characters
Measurement: 100 pages (65x80) = 520000 characters

The transmitted data bytes were measured here.

Results

Berkeley lpd resulted in the following ratios of transmitted data to user data

Print job, 1 page body text with Postscript	234%
Print job, 10 pages body text with Postscript	130%
Print job, 100 pages body text with Postscript	122%

Print job, 1 page body text with HP PCL	160%
Print job, 10 pages body text with HP PCL	112%
Print job, 100 pages body text with HP PCL	110%

Print job, 1 page body text with SWIN	145%
Print job, 10 pages body text with SWIN	110%
Print job, 100 pages body text with SWIN	108%

SAPlpd resulted in the following ratios of transmitted data to user data

Print job, 1 page body text with Postscript	146%
Print job, 10 pages body text with Postscript	53%
Print job, 100 pages body text with Postscript	38%

Print job, 1 page body text with HP PCL	133%
Print job, 10 pages body text with HP PCL	50%
Print job, 100 pages body text with HP PCL	36%

Print job, 1 page body text with SWIN	128%
Print job, 10 pages body text with SWIN	47%
Print job, 100 pages body text with SWIN	35%

Frontend printing resulted in the following ratios of transmitted data to user data

Print job, 1 page body text with SWIN	376%
Print job, 10 pages body text with SWIN	235%
Print job, 100 pages body text with SWIN	258%