speed

ATTACHED: comparison of download waterfall before and after applying these hacks to my page.

Hi Team,

I wanted to share with you some techniques I've discovered over the last couple of days that seem to have a noticeable impact on download times. These are especially applicable to pages with large CSS files on first load, as once things get cached it doesn't make much difference.

It seems that with the growing Drupal user base, the problems with CMS (flexibility) and speed (non functional requirement) seems to be growing. We made a SAAS where one can install a small module (http://drupal.org/project/paas) and the buy a service where we scan your site. We do this on 100+ metrics and in an academic way (lots of "r" graphs, no gradients :-)

We measure your site from a selectable EC2 location and really test the site insite out and outside in. We dont just say "use parallel downloading", we measure how many host you need to get the most benefit, no yslow wisdom! :-)

Hello Team,

I have heard about database latency, when we have a seperate DB server. But a lot of people are using them nevertheless.

We can also have a single server with 16 GB or More then that RAM.

Some says dedicated server are always good some says more RAM on same server, I am confused.

Which one would be better? Please suggest.

Thank You.

Hi Everyone.

Very important topics have been discussed in the high performance group in the past and i would like to request all members to help make a list of important factors, which can offer performance boost for Drupal website. I will list a few i know work very nicely. Please do add your suggestions.

1. Use BOOST module to serve static content to users
2. Cache router can also be used for caching.
3. If you are on dedicated hosting, use LiteSpeed Webserver. Even there are some companies offering LiteSpeed hosting in shared environment.

While searching for how to integrate drupal with a "light" webserver called Hiawatha, I find this comment intriguing:

http://www.hiawatha-webserver.org/forum/topic/222

In one of the comments, Hugo Leisink says that he doesn't recommend drupal because of the "ugly" code.
He says "On a simple page that showed a list of 15 items, Drupal needed more that 400 database queries."

I know drupal does these queries so it can do logging and for other status information, but I think you can disable these logging features.

To make it easier to find information on these matters, I've started the new handbook page Drupal caching, speed and performance in the Tutorials section of drupal.org. It's just a guide to information and tutorials on optimizing Drupal's performance, speed, and scalability, with an initial listing of fourteen resources. If anything is missing, you may edit it to add other useful Drupal performance resources.

As seen on Planet Drupal, the recent article Improving Drupal's Performance with the Boost Module for the UN's Millennium Campaign (October 23rd, 2008) describes how a Drupal site successfully managed a very high traffic situation. They achieved this with "one enormously helpful Drupal module, called Boost", and "some fine tuning".

I just wrote a new article called How I Survived a 2300% Traffic Increase With Drupal. It documents my experience with hitting the front page of Reddit, and some of the steps I took to make sure my site stayed online.

I also included some nice traffic graphs and statistics. If you ever wondered what kind of numbers to expect from a front-page link on a big social news site, be sure to check it out.

--
John Forsythe

Hi,

We need our www.projectstars.com site to run a lot faster than it is at the moment and wondered if anyone with experience could advise us on a good option for upgrading from our current virtual server. We're using memcache, block caching, and general drupal caching but as most of the site is dynamic I'm guessing that having a fast database server would help.

Hi

I am thinking about tuning my vote_up_down based site into a bit more digg-like site. On Digg 50 votes per hour can outperform 500 votes per two weeks - it allows "hot" news to be pushed to the front page earlier. Can similar things be achieved with the help of voting_api?

1. I am thinking about attaching to every node, not only a "vote" counter, but also kind'a "internal priority" counter, that would be increased by user votes and decreased by time therefore making old votes less valuable.