here are terms to google for:
caching query analysis profiling logfile analysis website statistics database indexes javascript/CSS file compression "rails go vroom"
Good luck!
Hi shiva,
this is what i did, which i feel might help you,
1. First i installed ruby-prof(very good tool) gem, and i profiled my requests. i found that my static database table queries are consuming 20%-30% of my response time. so planning to use acts_as_cached (which while spiking showed 18-20% boost)
2. And i am using gettext for internationalization which was consuming around 20% of response time, as i thought, this was unnecessary for default(English) language. So, i wrote a wrapper for gettext, which straight away gave 20% boost.
3. And just did some code cleanup.(refactoring).
4. And find tuned query to for faster response time.
5. And implemented file level caching logic for tables, which are very huge(partitioned table).(this gave 21%-25%)
As gene said planning to implement javascript/stylesheet/image compression/caching, and mongrel_rails with apache will be better approach for production deployment. and i am still in rails 1.1.6, so upgrading to rails 1.2.3 is on the way.(don't know this may give some boost, may be)
one more thing is if you are using oracle, you can cursor_sharing, prefetch_limit values in database.yml file to 'similar' and 500(suitable for me).
and One more thing is Do some google on GC(Garbage Collector) of ruby, which you can fine tune according to your application, surely the above approach will give some boost.
Note: all these results are with respect to 10 users accessing my application concurrently.
Good Luck,
Hi shiva,
i guess this will help for sure,