I use xFPC on most of my sites and it occured to me that once the xFPC cache is populated every page of the website is available at /core/cache/xfpc/ named as a sha1 hash of the page url.
If the cache files were named to match the page url and had an .html extension eg.
- category-product-item1.html for www.domain.co.uk/category/product/item1.html
url rewriting could be used to sidestep the PHP engine and return the page more quickly.
Obviously this would require some adjustments to xFPC but I ran some quick tests on the time taken to return the page:
- 60ms .html
- 160ms xFPC active
- 210ms xFPC inactive
And it looks like there is a saving of around 100ms to be had.
The major drawback I see with this method is there is no way to determine whether a user is logged into the Manager to sidestep the cache.
Are there any potential issues I missed and is there any merit in looking at this further?
It appears my brain needs a restart. I knew Statcache existed but my brain was telling me it works like getCache ie. caching elements rather than full pages.
I'll give Statcache a try.
Thanks.