Buffering in php

ade234uk

Today I found out about buffering in php. Now I have seen this, another new door may be opening?

Ok but does it have advantages?

Lets say I have a website, that has

10,000 products
500,000 page views a day
600 unique visitors a day

I suppose you could pull information from the database about the products and then buffer it?

And then just output it when you need to, when someone revisits the same product?

If you buffered the product info, would you need query the database again after this point?

Would it take the load off the database?

bradgrafelman

Are you talking about output buffering e.g. [man]ob_start/man ? If so, that data is lost once the script has finished.

The only other way I could think of "buffering" data retrieved from a database would be to write that data to a file, and then only update the file with data from the DB at a certain time interval.

Though, with the numbers you provided, I doubt 500k page views a day (unless they're highly concentrated at one specific time interval) is really going to break a DB (again, assuming you don't have a little Pentium 2 machine with 64MB of RAM or something like that).

ade234uk

yes I was talking about
ob_start()

MarkR

Output buffering doesn't do caching of any kind- it only affects a single request.

If you really have performance issues and think you can solve them by caching commonly visited pages (of course, do some profiling to determine this first), then you could possibly use the cache in eaccelerator.

Mark

Weedpacket

Buffering is not the same as caching, but you can take the contents of a buffer and cache it (the PEAR::Cache_Lite package uses the ob_* functions to do exactly this)