SAP Library - Find your answers

SAP Workload Analysis: Average Database Time per Dialog Step (ms)

A large portion of the "dispatched" time in a dialog is due to the reading and changing of data in the database, This field stores the time required for processing logical database calls (calls to the SAP database interface).

Various factors influence the logical database access time. Read requests can be responded to either by the database buffer or the fast local SAP buffer. The buffer efficiency, the number of the required accesses, and a high number of DB change accesses influence the total access time.

The database access time is measured in the database interface in the application server. This is why this also includes part of the application server CPU time and the network transfer time.

SAP Workload Analysis: Average Roll In Time per Dialog Step (ms)


Time used during a dialog step for rolling in the user context from the roll area into the local process memory.


During roll-in and roll-out operations, user-specific data from the roll buffer of shared memory or the roll file is loaded into the work process area and exported to the roll buffer or roll file.

SAP Workload Analysis: Average Response Time per Dialog Step (ms)

Response time per dialog step

The response time of a dialog step is the time required for requesting the dialog from the dispatcher work process to the processing of the dialog through the ending of the dialog in the dispatcher and the transfer of the data to the presentation server.

This also includes the time used for "Roundtrips" to transfer data from the SAP R/3 frontend to the application server and back. This time is recorded as roll wait time on the application server, while it is displayed as GUI time on the frontend. For inefficient networks, such as WAN connections, this time can contribute significantly to the response time, although it uses no resources on the application server, as the context is rolled out and the work process is released.

The response time is usually split into wait time and "dispatched" time. The SAP response time is made up of the following components:

Response time = wait time + "dispatched" time

"Dispatched" time

= Generation times during the runtime

+ Load times for programs, screens, and GUI interfaces

+ Roll times for rolling in work data

+ ABAP processing times

+ Database time

+ Enqueue time for logical SAP locks

+ Roll wait time (not including task types RFC/CPIC/ALE).

The CPU time is not an additive component of the response time, but rather the sum of the CPU time used by the individual components. The CPU time is an independent additional piece of response time information.

SAP Workload Analysis: Average CPU Time per Dialog Step (ms)

The CPU of the application server is used during a dialog step for processing. This includes, for example, loading and generating objects or database queries as well as processing ABAP commands.

The operating system determines the CPU time. At the end of a transaction step, the SAP R/3 work process queries the operating system for CPU time used. The CPU time is not an additive part of the response time (unlike wait, roll in, load, and database times).

SAP Workload Analysis: Average CPIC / RFC Time per Dialog Step (ms)

Average CPIC/RFC time per dialog step

For task types RFC/CPIC/ALE, this is only the time used in the CPIC/RFC interface.

For all other task types, the RFC processing time is added as well.

Step by step process on how to upgrade the kernel

"Updating the Kernel: Make sure you have at least 200MB of free disk space on the SAP instance before doing a Kernel update. The following steps illustrate how to update the Kernel on a central instance:

1. Download the latest Kernel patch into a temporary directory in your system or network share.

2. Unpack the patch using SAPCAR.exe. First, change to the directory you downloaded the patch into and then run SAPCAR –xvf .

3. Stop the SAP central instance and the SAP services.

4. Back up all existing files in the Kernel directory (usr/sap/sid/sys/exe/run). This way, you can easily recover the old Kernel in case of problems with the new one.

5. Copy the new Kernel files into the run directory.