Factors impacting page load speed

[Former Member]

We are experiencing some differences in page load speed (primarily on Reports) across different users and user types at my client's site. Here is our scenario:

We have connected to LDAP data via NWA and have assigned LDAP users to UME groups. These UME groups are then assigned to Application Groups in SSM Admin, and these are subsequently tied to specific Contexts.

The user in question and I are both logging in from within the company network but are in different physical locations (different states). When I VPN into the network and access SSM remotely the page load speed is the same as it is when I am on site.

Another LDAP user at the same location as the user in question experienced normal page load speed.

The user in question experienced slow page load speed from a shared PC in a conference room.

To see if we can link load speed differences to this specific LDAP user account I had the user log in as pipadmin. The load speed was much faster when logged in as pipadmin.

When I user my own LDAP user ID and compare it to the speed of pipadmin, it's the same.

Questions:

1. Should system access as a UME user be any different (faster/slower) than as an LDAP user?

2. Our administrative users are assigned to multiple UME groups. Could this impact system speed?

3. Given the circumstances described above, is it possible that the location of the user in question (i.e., the LAN at his office) has anything to do with system speed?

4. We are using ODBO to connect to BW. Would the system be faster if we switched to BAPI?

Thanks!

Laura

[former_member84747]

Hi Laura,

I'll try and answer your questions.

1. There should be no difference in access speed for a user coming from LDAP versus one which is defined in UME. Internally, SSM treats all users in exactly the same way. In fact, there is no difference between an LDAP user and one from UME. We create one overall user array internally made up of users from both sources simply passed by CE.

2. The number of groups a user is assigned to should have no meaningful impact on performance. There could be a millisecond or so's difference somewhere internally, but it shouldn't be noticeable by the users.

3. This seems the most likely explanation for performance. My suggestion would be to use a tool like HTTP Watch or Fiddler to examine the load times for opening specific parts of the page, and compare where the delays are coming from.

4. I don't know that there is a significant performance improvement with BAPI, but it is a much more stable connectivity piece and does not require that the SAP front end be installed in order to work. Also, we are no longer testing ODBO moving forward, because some versions of the ODBO driver work while others do not, so I would recommend moving to BAPI for all those different reasons.

Hope this helps,

Robert