This content has been marked as final. Show 9 replies
There are so many factors that will influence thet page load time but a well designed, well implemented, well configured webcenter portal application should be able to serve a page below 500ms.
Remember that WebCenter is the top of a complete stack and there are many points of failure that can cause delays: firewall, http server, load balancer, access manager (LDAP), network latency, hardware, database,...
All of these components need to work together in a proper way to serve a good response time.
On my implementation of Spaces PS5, using the Oracle IDAM suite, we see ~10-30 second logins and ~4-6 second page response times.
When we remove all of our content (task flows such as content presenters and doc list viewers) from the page, page responses improve to ~1 second.
Anyone else struggle with login or page response times in Webcenter Spaces?
Have a look for the performance issues,
Thanks for the links. These are good resources.
The recommendation of 5 task flows is helpful. It makes sense but I hadn't heard a maximum recommended. The number of task flows used on pages varies but there are 4 - 8 task flows per page. A typical page will have:
- Doc List Viewer
- 2 - 5 content presenters querying WCC
We are in the process of configuring coherence to reduce the impact of the content presenters. We haven't dug into JOC, still using whatever comes out of the box.
Are there down sides to having coherence caches valid for extended periods of time (i.e. 30 days)? Does the cache know when content is updated so it doesn't provide old content?
Any other ideas on how to improve performance for pages that aren't previously cached?
2 to 5 content presenters on a single page is to much. You need to try to consolidate the content so you can minimize the number of content presenters on a page.
Most of the time the problem is in the content model. Even complex content models can be put in a single data file by using lists and so on. This way you can minimize the CP's to one or two per page.
When using coherence you should always set the expiry-delay for the ContentNodeCaches to 0 which means that coherence will never invalidate the cache!
The reason behind this is that the connection to UCM implements a content sweeper which will invalidate the cache items that have been changed. In the Enterprise Manager you have the field "Cache Invalidation Interval". This will determine the interval in which a service from UCM is called to check which items have been updated. This will also notify coherence and update the nodes with the new version.
That's why the expiry delay needs to set to 0 in coherence.
Also set the high units to 80% of your entire repository that is used on the top pages. This can easily be 10000 without problem.
I'm planning on writing a blog post about these configurations later on.
The JOC configuration is OOTB but you need to run a script to enable it. It is recommended when you use WC Spaces.
Information can be found here: http://docs.oracle.com/cd/E23943_01/core.1111/e12037/extend_wc.htm#CHDIFEJH