Hi there,

When i surf dnn page, some times it hangs. Some times it take too long time to open light box (for login, for example). I saw some sites that used dnn, but they are fast in my opinion.

It's worth to mention that my site is on local machine.

 

Thanks.

If these database scripts speed up DNN, why are they not included in the core DNN database schema? Why is this left to an external party to provide database indexing modifications? Database design in fundamental - it's the base upon which everything else is built!! Indexing is essential, but sometimes, indexing is a mechanism to assist bad database design...as I believe DNN has.

Sebastian, I like what you've done - don't get me wrong.. DNN has been in desperate need of of a properly designed and indexed database for as long as I can remember, but why doesn't the DNN mothership do this as part of the natural evolution of the database?

IMO, the DNN eco-system needs a database schema certification process. So many vendors create tables, views, stored procedures etc, that have no naming convention, poor structure and no indexing, or referential database integrity.

One of the most important rules when dealing with optimization is to never change anything you can't measure - whilst we've accepted many (in Sebastian's case around 70) optimization enhancements, most of those could be accepted as they addressed clearly inefficient stored procedures, or were missed indexes where indexes were sure to provide a benefit. However other enhancements need to be measured and metrics produced, to ensure we do not introduce bugs or inefficiencies e.g. a change may speed up selects but negatively effect changes (i.e. insert/update/delete). This includes not just the time to perform the queries but any resultant actions such as rebuilding indexes etc. At the minute we don't have this data, so need to measure these items (with a number of different test sites to ensure the results are consistent) before we can consider adding many of these proposed optimizations.

We've recently rebuilt our performance testing setup so we can accurately benchmark performance and what is occurring, particularly in reference to common scenarios e.g. browsing anonymously, logging in, using the common admin functions (we also have versions with substantial data loads). We're using this to evaluate where performance can be improved - this goes beyond the database, and covers all aspects (script, page size, code, database etc.). The results of this are very interesting e.g. in a lot of cases we're seeing that two thirds of page load time is caused by DOM interaction/scripts - so changes to reduce/combine script or simplify the DOM can boost page loads a lot, so at the minute focussing on this provides the most short-term benefit (we're also working on reducing start-up overhead and memory usage, as well as database optimization). We'll continue working with this setup to identify the bottlenecks and costly operations, and in time I'm sure we'll hit the other items Sebastian raised.

---

Buy the new Professional DNN7: Open Source .NET CMS Platform book Amazon US